Skip to main content

Full text of "Restoration of American Chesnut to Forest Lands: Proceedings of a Conference and Workshop Held May 4-6, 2004 at The North Carolina Arboretum"

See other formats


National Park Service 

U.S. Department of the Interior 



29.89:NPS5/NCR/CUE/NRR-2006/O1 



National Capital Region 
Center for Urban Ecology 




3 1604 017 801 756 






Restoration of 

American Ciiestnut 

to Forest Lands 



Proceedings of a Conference and Workshop 

Held May 4-6, 2004 

at The North Carolina Arboretum 

Edited by Kim C. Steiner 
and John E. Carlson 



Natural Resources Report NPS/NCR/CUE/NRR — 2006/01 



RESTORATION OF AMERICAN CHESTNUT 

TO FOREST LANDS 



Proceedings of a conference and workshop 

held at 

The North Carolina Arboretum 

Asheville, North Carolina, U. S. A. 

May 4-6, 2004 

Natural Resources Report NPS/NCR/CUE/NRR - 2006/001 



Organized and Edited by: 

Kim C. Steiner, Professor of Forest Biology 

and 
John E. Carlson, Associate Professor of Molecular Genetics 

School of Forest Resources 
The Pennsylvania State University 



Sponsored by: 

USDl National Park Service 

National Capital Region 

Chesapeake Watershed Cooperative Ecosystem Studies Unit 

and 
Southern Appalachian Mountains Cooperative Ecosystem Studies Unit 



May 2006 

us. Department of the Interior 
National Park Service 
National Capital Region 
Center for Urban Ecology 



Natural Resources Reports are the designated medium for information on technologies and resource 
management methods: "how to" resource management papers; proceedings of resource 
management workshops or conferences: and natural resource program descriptions and resource 
action plans. 

Views and conclusions in this report are those of the authors and do not necessarily reflect policies 
of the National Park Service. Mention of trade names or commercial products does not constitute 
endorsement or recommendation for use by the National Park Service. 

This report was accomplished under Cooperative Agreement 1443CA309701200. Task Agreement 
Number T-3097-0 1-002 with assistance from the National Park Service. The statements, findings, 
conclusions, recommendations, and data in this report are solely those of the athors and do not 
necessarily reflect the views of the U.S. Department of the Interior. National Park Service 

Printed copies of this report were produced in a limited quantity and, as long as the supply lasts, 
may be obtained by sending a request to the address on the back cover. When original quantities 
are exhausted, copies may be requested from the NPS Technical Information Center (TIC), Denver 
Service Center, PO Box 25287. Denver, CO 80225-0287. A copy charge may be involved. To 
order from TIC, refer to document D-66. 

This report may also be available as a downloadable portable document format file from the Internet 
at http://chestnut.cas.psu.edu/nps.htm 

Acknowledgements: 

Funding for this conference was provided by the United States Department of The 
Interior. National Park Service, through the Chesapeake Watershed Cooperative 
Ecosystem Studies Unit located in Frostburg. Mar> land. The editors would like to 
acknowledge the very considerable support provided by Dr. William Lellis in 
arranging the funding for this conference and in assisting with its organization and 
execution. We also wish to thank Ms. Sara Fitzsimmons for her expert help with 
editing and formatting the proceedings. 

Citation: 

Steiner, K. C. and Carlson. J. E. eds. 2006. Restoration of American Chestnut To Forest 
Lands - Proceedings of a Conference and Workshop. May 4-6. 2004. The North Carolina 
Arboretum. Natural Resources Report NPS/NCR/CUE/NRR - 2006/001. National Park 
Service. Washington. DC. 

Front Cover: 

The figure on the front cover depicts the core of the natural range of American 
chestnut (as delineated by Elbert Little) superimposed on a map of the present 
occurrence of forest in the United States (http://nationalatlas.gov/mld/foresti.html) 
and Canada (http://geogratis.cgdi.gc.ca/clf/en). Small, outlying populations of the 
original American chestnut range are not shown. 



NPS D-66 May 2006 



^ TABLE OF CONTENTS 

Preface v 

List of Conference Participants vi 

Welcome and Introduction 

James Sheralci 1 

National Park Service Management Policy Guidance for Restoration of American Chestnut 
{Castanea dentata) to National Park System Units 

John Dennis 3 

Ecology 

Historical Ecology of American Chestnut {Castanea dentata) 

Emily Russell Southgate 13 

Forest Health Impacts of the Loss of American Chestnut (transcript of presentation) 

Steve Oak 21 

Current Status of Chestnut in Eastern US Forests 

William McWilliams, Tonya Lister, Elizabeth LaPoint, Anita Rose, and John Vissage 31 

Chestnut and Wildlife 

Frederick Paillet 41 

Historical Significance of American Chestnut to Appalachian Culture and Ecology 

Donald Davis 53 

Blight Resistance Technology 

The Backcross Breeding Program of The American Chestnut Foundation 

Fred He bard 61 

Blight Resistance Technology: Transgenic Approaches 

William Powell, Scott Merkle, Haiying Liang, and Charles Maynard 79 

Hypovirulence: Use and Limitations as a Chestnut Blight Biological Control 

William MacDonald and Mark Double 87 

Integrated Use of Resistance, Hypovirulence, and Forest Management to Control Blight on 
American Chestnut 

Gary Griffin, John Elkins, Dave McCurdy, and L ucille Griffin 97 



III 



Genetic Issues — ^=^ 

Genetic Structure of American Chestnut Populations Based upon Neutral DNA Markers ^— ^ 
Thomas Kiibisiak and James Roberds 1 09 

Regional Adaptation in American Chestnut =^=-^ 
Kim Steiner 123 

Rate of Recovery of the American Chestnut Phenotype Through Backcross Breeding of Hybrid 
Trees 

Matthew Diskin and Kim Steiner 1 29 

Selection for Chinese vs. American Genetic Material in Blight-Resistant Backcross Progeny 
using Genomic DNA 

Song Liu and John Carlson 133 

Biological Confinement of Genetically Engineered Organisms 

John Carlson 151 

National Park Service Policies - Highlights From a Workshop on Genetically Modified 
Organisms in Park Lands 

William Lellis ! 159 

Practical Issues 

Planting Trials of American Chestnut in Central Appalachian Forests 

Tim Phelps, Kim Steiner, Chien-Chih Chen, and James Zaczek 161 

Planting Trials with American Chestnut in Southern Appalachian Forests 

David Loft is 167 

Ecosystem Restoration and Federal Land Policy: Reexamination in Light of the American 
Chestnut Restoration Effort 

Robert McKinstry, Jr. 1 73 

Feasibility of Large-Scale Reintroduction of Chestnut to National Park Service Lands 

Scott Schlarbaum, Sunshine Brosi, and Sandra Anagnostakis 195 

Effects of Past Land Use and Initial Treatment on Castanea dentata Seedlings 



f 



Jennifer Hewitt, Albert Meier, John Starnes, Priscilla Hamilton, and Charles Rhoades .... 203 

Potential Extent of American Chestnut Restoration within the National Park System 

William Lellis 211 

Workshoi) 

Summary of Facilitated Workshop of Restoration of Chestnut to National Park System Lands 

James Finley and Kim Steiner 227 



IV 



V PREFACE 

The papers in this document were presented at a conference and workshop that was held at The 
North CaroMna Arboretum in Asheville, NC, May 4-6, 2004. The purpose of the conference was 
to discuss issues surrounding the restoration of American chestnut to forest lands. The audience 
members were primarily employees of USDI National Park Service (NPS), and interest focused 
on the question of restoring chestnut to NPS lands, but most presentations were selected to 
address restoration from the broadest perspective possible. The organizers of the conference were 
Drs. Kim Steiner and John Carlson of the School of Forest Resources at The Pennsylvania State 
University, and the sponsors were the Chesapeake Watershed Cooperative Ecosystem Studies 
Unit (CESU) (NPS), the Southern Appalachian Mountains CESU (NPS), and The Pennsylvania 
State University. 

The conference covered the current status of chestnut blight research and objectives, 
opportunities, and potential directions for American chestnut restoration programs on NPS lands. 
Topics discussed at the meeting included policy issues, the current status of chestnut, chestnut 
ecology, breeding programs, blight resistance technologies, genetic issues, potential impacts on 
forest ecology, design of restoration programs, and knowledge gaps related to restoration within 
the National Park System. The conference ended with half-day workshop facilitated by Dr. 
James Finley of the School of Forest Resources at The Pennsylvania State University. Attendees 
remarked that the scope and quality of presentations established the meeting as a benchmark 
event in the history of chestnut restoration. As a result of the meeting, a summary of issues and 
recommendations for National Park Service administrators is being prepared. This collection of 
papers represents the most comprehensive and current information available at this time on the 
biology of American chestnut and the blight fungus and the potential for restoring chestnut to its 
native range. 



John E. Carlson and Kim C. Steiner, August 4, 2005 



LIST OF CONFERENCE PARTICIPANTS 



Ray Albriglit 

Southern Appalachian Mtns. CESU 

University of Tennessee 

274 Ellington Plant Science Building 

Knoxville,TN 37966 

Ray_Albright@nps.gov 

865-974-8443^ 



John Bellemore 

USDA Forest Ser\ ice 

G. Washington & Jefferson National Forests 

5162 Valley Pointe Parkway 

Roanoke, V A 24019 

jbellemore@fs.fed.us 

540-265-5150 



Mark Alexander 

University of Tennessee, Chattanooga 

Mark.dxmf27)hotmail.com 

Stephen Alexander 

University of Tennessee, Chattanooga 

morganfreemore@hotniail.com 

Kristen Allen 

National Park Service 

Richmond National Battlefield Park 

3215 E. Broad Street 

Richmond, VA 23223 

kristen_allen@nps.gov 

804-795-5019' 

Mary Willeford Bair 

National Park Service 

Shenandoah National Park 

3655 US HWY 211 

East Luray, VA 22835 

mary willeford_bair@nps.gov 

540-999-3490 

Clarissa Balbalian 

Mississippi State University 

Extension Service 

P.O. Box 9655, Bost Rm. 9 

Mississippi State, MS 39762-9655 

cbalbali@ext.msstate.edu 

662-325-2146 

Jenny Bccler 

National Park Service 

Cumberland Gap National Historical Park 

P.O.Box 1848 

Middlcsboro, KY 40965 

jenny bceler@nps.gov 

606-246-1113 



Paul Berrang 
USDA Forest Service 
626 E. Wisconsin Ave 
Milwaukee, Wl 53202 
pberrang@fs.fed.us 
414-297^-3569 

John Blanton 
USDA Forest Service 
Southern Research Station 
P.O. Box 2750 
Asheville, NC 28802 
iblanton@fs.fed.us 
828-257-4248 

Tom Blount 

National Park Service 

Big South Fork National Recreation Area 

Oneida, TN 

tom_blount@nps.gov 

Wayne Bow man 

Virginia Department of Forestry 

900 Natural Resources Drive 

Charlottesville, VA 22903 

bovvmanw^aidof.state.va.us 

434-977-1375x3331 

Amanda Brennan 

National Park Service, Blue Ridge Parkway 

almbrcnnan@care2.com 

John Carlson 

Penn State University 

School of Forest Resources 

304 Wartik Lab 

University Park. PA 16802 

jecl6@psu.edu 

814-863-7561 



VI 



Brian Carlstrom 
National Park Service 
Prince William Forest Park 
18100 Park Headquarters Rd. 
Triangle. V A 22 172 
Brian_Carlstrom@nps.gov 
703-221-3329 

Marshal T. Case 

The American Chestnut Foundation 

469 Main St.. Suite 1. P.O. Box 4044 

Bennington, VT 05201 

niarshalc@acf.org 

802-447-0110 



Donald Edward Davis 
Dalton State College 
213 N. College Dr 
Dalton, GA 30720 
ddavis@em.daltonstate.edu 
706-428-2928 

Mark DePoy 

National Park Service 

Mammoth Cave National Park 

P.O. Bo.v 7 

Mammoth Cave, KY 42259 

mark_depoy@nps.gov 

270-758-2141 



Barry Clinton 
USDA Forest Service 
Coweeth Hydrologic Lab 
3160CoweetaLabRd. 
Otto, NC 28763 
bclinton@fs.fed.us 
828-524-2128 X 124 



Matt Diskin 

Penn State University 

School of Forest Resources 

132 1/2 E. Prospect Ave. 

State College, PA 16802 

msaucel 999@yahoo.com 

814-237-8915 



Ries Collier 

National Park Service 

Cumberland Gap National Historical Park 

P.O.Box 1848 

Middlesboro, KY 40965 

ries_collier@nps.gov 

606-246-1110 

Benji Cornett 

The American Chestnut Foundation 

Meadowview Research Farms 

benji@acf.org 

J. Hill Craddock 

U. Tennessee at Chattanooga 

Biology & Environmental Sciences 

615 McCallie Ave. 

Chattanooga. TN 37403 

Hill-Craddock@utc.edu 

423-425-4643 

Barbara Crane 
USDA Forest Service 
Southern Region 
1720PeachtreeRd.NW 
Atlanta, GA 30309 
barbaracrane@fs.fed.us 
404-347-4039 



Coleman Doggett 

Duke University 

2 1 7 Rosecommon Lane 

Gary, NC 27511 

ncdoget@mindspring.com 

919-467-0551 

Greg Eckert 

National Park Service 

1201 Oakridge Drive, Suite 200 

Fort Collins, CO 80525 

greg_eckert@nps.gov 

970-225-3694 

Katherine Elliott 
USDA Forest Service 
Coweeta Hydrologic Laboratory 
3160CoweetaLabRd. 
Otto, NC 28763 
kelliott@fs.fed.us 
828-524-2128 X 110 

Songlin Fei 

Penn State University 

School of Forest Resources 

9 Ferguson Building 

University Park, PA 16802 

feisl@psu.edu 



Vll 



James Finley 

Perm State University 

School of Forest Resources 

7 Ferguson Building 

University Park, PA 16802 

jfinley@psu.edu 

814-863-0401 

Sara Fitzsimmons 
Penn State University 
210 Forest Resources Lab 
University Park, PA 16802 

sfO@psu.edu 
814-865-7228 

Sharon Friedman 
USDA Forest Service 
Washington Office 
Washington, DC 
sfriedman@fs.fed.us 
202-205-0939 

Manley Fuller 

Florida Wildlife Federation 

P.O. Box 6870 

Tallahassee, FL32314 

wildfed@aol.com 

850-656-7113 

Peter Gould 

Penn State University 

School of Forest Resources 

Forest Resources Lab 

University Park, PA 16802 

Pjgl69@psu.edu 

Gary Griffin 

Virginia Tech University 

Plant Path., Phys., and Weed Science 

Blacksburg.VA 24061 

gagriffUcT'vt.edu 

540-552-5943 

Brian Heath 

North Carolina Forest Service 

Asheville, NC 

brian.heath@ncmail.net 



Fred Hebard 

TACF Research Farms 

14005 Glenbrook Ave. 

Meadowview, VA 24361 

Fred@acforg 

276-944-463 1 

Jennifer Hewitt 
National Park Service 
Mammoth Cave National Park 
Jennifer Hewitt@partner.nps.gov 

Larry Hilaire 

National Park Sersice 

Delaware Water Gap Recreation Area 

294 Old Milford Rd. 

Milford, PA 18337 

larry_hilaire@nps.gov 

570-296-6952 x 27 

Hugh Irwin 

46 Haywood St., Suite 323 

Asheville. NC 28801 

hugh@safc.org 

828-252-9223^ 

Joseph James 

The American Chestnut Foundation 

Carolina's Chapter 

260 Steve Nix Rd. 

Seneca, SC 29678 

s4e4j4@bellsouth.net 

864-972-1122 

Brian Joyce 

Montrcat College 

Environmental Studies 

301 Morgan Science Building 

Montreal, NC 28757 

bjoyce@montreat.cdu 

828-654-8623 

John Karish 
National Park Service 
University Park, PA 
John Karish@nps.gov 
814-865-7974 



VIII 



Robert Kellison 

Institute of Forest Biotechnology 

15 T.W. Alexander Dr. P.O. Box 3399 

Research Triangle Pk, NC 27709-3399 

bob_kellison(S)ncbiotech.org 

919-549-8896 

Jennifer Knoepp 
USDA Forest Service 
jknoepp@fs.fed.us 

Thomas Kubisiak 
USDA Forest Service 
Southern Research Station 
23332 Highway 67 
Saucier, MS 39574-9344 
tkubisiak^fs.fed.us 
228-832-2747x213 

Jennifer Lee 
National Park Service 
Prince William Forest Park 
18100 Park Headquarters Rd. 
Triangle, VA 22 172 
jennifer_lee@nps.gov 
703-221-3406 

BillLellis 

National Park Service 
Chesapeake Watershed CESU 
UMCES, Appalachian Laboratory 
301 BraddockRoad 
Frostburg,MD 21532 
wlellis@al.umces.edu 
301-689-7108 

Song Liu 

Penn State University 

304 Wartik Lab 

University Park, PA 16802 

szll 10@psu.edu 

814-235-9428 

David Loftis 

USDA Forest Service 

Southern Research Station 

Asheville, NC 

dloftis@fs.fed.us 

828-667-5261 x 115 



Rebecca Loncosky 
National Park Service 
Catoctin Mountain Park 
6602 Foxville Rd. 
Thurmont, MD21788 
becky_loncosky@nps.gov 
301-416-0536 

William MacDonald 

West Virginia University 

College of Agriculture and Forestry 

401 Brooks Hall, P.O. Box 6058 

Morgantown, WV 26506-6058 

Bill.MacDonald@mail.wvu.edu 

304-293-3911 x2236 

Russ MacFarlane 
USDA Forest Service 
rmacfarlane@fs.fed.us 

Jim Maldox 

Tennessee Valley Authority 

Muscle Shoals. AL 

Rex Mann 

USDA Forest Service 

Daniel Boone National Forest 

1 700 Bypass Rd. 

Winchester, KY 40391 

rmann@fs.fed.us 

859-744-7086 

Susan McCord 

Institute of Forest Biotechnology 

15 T.W. Alexander Dr. P.O. Box 3399 

Research Triangle Pk, NC 27709-3399 

susan_mccord@ncbiotech.org 

919-549-8889^ 

Robert McKinstry 
Penn State University 
102 Ferguson Building 
University Park, PA 16802 
rbmlO(a)psu.edu 
814-865-9390 



Chris McNeilly 

National Park Service 

Kings Mountain National Military Park 

2625 Park Road 

Blacksburg, SC 29702 

chris_mcneilly@nps.gov 

864-936-7921 

Will McWilliams 
USDA Forest Service 
Northeastern Research Station 
1 1 Campus Blvd., Suite 200 
Newtown Square, PA 19073 
vvmcwilliamsfSifs.fed.us 
610-557-4050^ 

Albert J. Meier 
Department of Biology 
Western Kentucky University 
Albert.meieri@wku.edu 

Dan Miller 

USDA Forest Service 

320 Green St. 

Athens. GA 30606 

dmiller03@fs.fed.us 

706-559-4247 

Steve Oak 

USDA Forest Service 

Southern Research Station 

P.O. Box 2680 

Asheville,NC 28802 

soak@fs.fed.us 

828-257-4322 

John Perez 

National Park Service 

New River Gorge National River 

P.O. Box 246 

Glen Jean, WV 25846 

John perez@nps.gov 

304-465-6537 

Timothy Phelps 
Penn State University 
School of Forest Resources 
210 I'orcst Resources Lab 
University Park, l>A 16802 
phelpst@psu.edu 
814-865-7228 



William Powell 

State University of New York 

1 Forestry Drive, 319 I Hick Hall 

Syracuse, NY 13210 

wapovvelI@esfedu 

3 1 5-470-6744 

Philip Pritchard 

The American Chestnut Foundation 

Southern Appalachian Regional Office 

One Oak Plaza, Suite 308 

Asheville,NC 28801 

pritchard@acforg 

828-281-0047 

Anita Rose 

USDA Forest Service 

anitarose@fs.fed.us 

Joe Schibig 

Volunteer State Community College 

1480 Nashville Pike 

Gallatin, TN 37066 

joe.schibig@volstate.edu 

615-452-8600x3270 

Scott Schlarbaum 

University of Tennesse 

Dep. of Forestry, Wildlife, and Fisheries 

Ellington Plant Science Building 

Knoxville.TN 37996 

tenntip@utk.edu 

865-974-7993 

Kent Schwarzkopf 

National Park Ser\'ice 

Appalachian National Scenic Trail 

High Street 

Civil War Stor> BIdg, 3rd Floor 

Harpers Ferry, WV 25425 

kent_schwarzkopf@nps.gov 

304-535-6767 

James Sherald 
National Park Service 
4598 MacArthur Blvd. NW 
Washington, DC 20007-4227 
Jim_Sherald@nps.gov 
202-342-1443x208 



Paul Sisco 

The American Chestnut Foundation 

Southern Appalachian Regional Office 

One Oak Plaza. Suite 308 

Asheville,NC 28801 

psisco^mindspring.com 

828-281-0047 



James Voigt 

National Park Service 

Catoctin Mountain Park 

6602 Foxville Rd. 

Thurmont,MD 21788 

cato_resource_management@nps.gov 

301-416-0536 



Emily Russell Southgate 
P.O. Box 642 
Middleburg, VA20118 
erussell(a)nac.net 
540-687-8291 

Brad Stanback 

8 1 Long Branch Road 

Canton, NC 28716 

ilexvert@mindspring.com 

828-646-9447 



Jason Walz 
National Park Service 

Geoff Wang 

Clemson University 

Dept. of Forestry and Natural Resources 

261 LehotskyHall 

Clemson, SC 29634-03 17 

gwang@clemson.edu 

864-656-4864 



Kim Steiner 

Penn State University 

School of Forest Resources 

213 Ferguson Building 

University Park, PA 16802 

steiner@psu.edu 

814-865-9351 

K. O. Summerville 

N.C. Division of Forest Resources (retired) 

Carolina's Chapter TACF 

1623 Kenbrook Drive 

Gamer. NC 27529 

kospcs@bellsouth.net 

919-772-7111 

Glenn Taylor 

National Park Service 

Great Smoky Mountains National Park 

Jack Torkelson 

Volunteer State Community College 

torkelsonj@bellsouth.net 

Chris Ulrey 

National Park Service, Blue Ridge Parkway 

199 Hemphill Knob Rd. 

Asheville, NC 28803 

chris_ulrey@nps.gov 

828-271-4779x271 



Keith Watson 

US Fish and Wildlife Service 

160 Zillicoa St., Suite D 

Asheville, NC 28801 

keith_watson@fws.gov 

828-350-8228 

Michele Webber 

Western Kentucky University 

Mammoth Cave National Park 

Michele.webber@contractor.nps.gov 

Stewart Winslow 

Millikew and Company 

Stewart.winslow@millikew.com 



XI 



Steiner, K. C. and Carlson, J. E, eds. 2006. Restoration of American Chestnut To Forest Lands - Proceedings of a 
.^ Conference and Workshop. May 4-6, 2004, The North Carolina Arboretum. Natural Resources Report 
NPS/NCR/CUE/NRR - 2006/001, National Park Service. Washington, DC. 

INTRODUCTION 

^_^ James L. Sherald 

.\^ Center for Urban Ecology, National Park Service 

4598 MacArthur Blvd., N.W., Washington DC 20007 USA (JimSheraldfSnps.gov) 



The American chestnut {Castanea dentata (Marsh.) Borkh.) once accounted for a quarter of the hardwood 
trees throughout the eastern deciduous forest and in some locations in the southern Appalachians, its 
density reached 70-85%. A rapid, dramatic decline in the species dominance began around 1900 with the 
introduction from the Orient of the chestnut blight fungus, Cryphouectria parasitica (MmnW) Barr. In 
spite of early efforts at eradication, within 30 years the fungus had spread throughout the chestnut's entire 
native range, which extends from Maine to Georgia, Alabama, and Mississippi and west to eastern 
Michigan and southern Illinois. By 1940 practically all of the chestnuts throughout the range were dead or 
infected. 

As the chestnut declined, the eastern deciduous forest dramatically changed. Other species filled the void, 
principally oak, hickory, and pine in the southern Appalachians creating the Mixed Oak, Oak-Pine, Mixed 
Mesophytic, and Oak-Hickory Forest associations and hemlock, sugar maple, and beech to the north in 
the Allegheny Mountains, which is primarily the Beech-Maple Forest association. While the chestnut has 
relinquished its dominant role, its legacy continues through its regenerative capacity to resprout from the 
roots of long infected trees. The American chestnut exists today largely as a clonal understory sapling or 
pole tree, rarely living longer than 1 to 40 years. However, some of these sprouted trees are able to set 
fruit before succumbing, but the new seedlings rapidly become infected. The importance of this sexual 
reproduction cannot be underestimated when thinking in tenns of evolutionary time. Someday, long in 
the future, there may be successful seedling recruitment. 

Few ecological disasters have generated as much interest as chestnut blight. Shortly after the disease was 
first recognized in New York in 1904, research endeavored to understand every aspect of the disease and 
its exotic causal agent. Over the years much has been learned. In recent decades significant progress has 
been made in several areas, including the selection and breeding for blight resistance and the discovery 
and enhancement of fungal hypovirulence. Hypovirulent strains of C parasitica contain infectious 
cytoplasmic viruses that reduce the ability of the pathoge.i to cause cankers. Much of this research has 
captured the interest of the public renewing hope that this iconic American species will eventually 
reappear in the eastern deciduous forest. The National Park Service, which manages many parks 
throughout the former chestnut domain, will undoubtedly be expected to engage these advances and fulfill 
this dream. While these developments may still be several or many years away from practical 
application, the prognosis for eventually managing chestnut blight is promising. Consequently, the 
National Park Service must begin to fully understand the promise and pitfalls of these advances and to 
explore the limitations and consequences of restoration. 

This series of presentations and the following discussion are intended to assist the National Park Service 
in fulfilling three objectives: 

Our first objective is to develop a comprehensive understanding of all the science and technology that 
hold significant promise for restoring the American chestnut. Hopefully, through these presentations we 
will be able to assess the feasibility and potential for success as well as the long-term consequences these 
advances could impose on the ecosystem. The significant areas of interest include biocontrol through the 
use of hypovirulence, the selection and breeding of naturally-occurring putative resistant American 
chestnuts, back crossing the American chestnut with the resistant Chinese chestnut, and transgenic 



approaches to enhance host resistance and to debihtate the pathogen, as well as the potential for ^, 

combining these technologies into an integrated program. ~ 

Our second objective is to define our goals for American chestnut restoration. We recognize that these 
technologies promise a range of restoration possibilities, from the minimal establishment of 
demonstration plantings to the incorporation of resistant selections and biocontrol agents into major __^- 
reforestation projects. While these technological advances are driving our immediate interest in 
restoration, our decisions must also be guided by an understanding and appreciation of the ecological 
consequences posed by restoration choices. The National Park Ser\ ice must ha\e a thorough discussion 
as to whether our restoration goals and the technologies we select to achieve these goals are compatible 
with our policies, management objectives and most importantly our resources. .^^ 

Based on the status and feasibility of the technology and our restoration goals, the third objective is to 
prescribe how the Service and the parks should proceed. What are the acceptable choices today, what 
promising technologies will we endorse in the future, what policy issues must we address, what research 
do we believe is still necessary on unanswered questions or issues, and how can the National Park Ser\ ice 
assist? The implementation of some technology, such as the use of transgenic chestnuts or bioengineered 
hypovirulent strains for biocontrol, may necessitate policy decisions. Decisions relating to transgenic 
organisms are also relevant to other restoration and management issues affecting the Ser\ ice and are 
under discussion now. Other approaches, such as the planting of hybrids may soon be a\ailable. 
However, this option, like all tree planting efforts, has long-temi consequences and the decision to 
proceed should be well founded. We may decide that the research findings are premature or inconclusive, 
the long-term prospects uncertain, and additional study is necessary before we begin to engage in large 
scale restoration programs. While understandably, parks may differ in their restoration objectives: their 
decision process must be consistent and based on the best available science. The visiting public has 
always been interested in the ecological and cultural heritage of the American chestnut in their National 
Parks. Consequently, the National Parks will present the most visible and critiqued application of 
chestnut restoration technology. The public fully and rightfully expects us to understand and support the 
decisions we make. 



Steiner. K. C. and Carlson, J. E, eds. 2006. Restoration of American Chestnut To Forest Lands - Proceedings of a 
£i^^^ Conference and Workshop. May 4-6, 2004, The North CaroHna Arboretum. Natural Resources Report 
NPS/NCR/CUE/NRR - 2006/00 1 , National Park Service. Washington, DC. 



I, 



NATIONAL PARK SERVICE MANAGEMENT POLICY GUIDANCE FOR RESTORATION OF 

AMERICAN CHESTNUT TO NATIONAL PARK SYSTEM UNITS 

John Dennis 
National Park Service, 1849 C Street, N.W., Washington, DC 20240, USA (john_dennis@nps.gov) 

Abstract: The National Park Service's Management Policies 2001 provide clear guidance for decisions 
regarding management of the nearly extirpated American chestnut (Casfanea deutata). Restoration is 
appropriate and may involve active planting, cross breeding, and genetic engineering using genotypes 
from areas within and outside the parks. Restoration must be based on science, analyzed through 
environmental compliance processes, include the public, and consider actively involving partners. 

Keywords: management policies / restoration / American chestnut / blight / exotic species 

INTRODUCTION 

Law, policy, philosophy, and science contribute to decision-making about whether or not to attempt to 
restore the American chestnut {Castauea dentata) to one or more units of the National Park System. The 
following discussion addresses these components of decision-making by examining specific provisions 
and then identifying possible pathways for applying them to management actions. 

EXCERPTS FROM NATIONAL PARK SERVICE MANAGEMENT POLICIES 2001 

The National Park Service develops and publishes the Management Policies to interpret the many laws 
that authorize and direct the purposes, uses, and management of lands incorporated into the National Park 
System and to fill in details not specifically addressed in the laws. The Service's Management Policies 
2001 provide both general and specific guidance regarding management of park natural resources. The 
statements provided in this section are taken directly from, or paraphrase, selected entries in the 
Management Policies 2001. 

Natural Conditions 

In its most general terms, the Management Policies 2001 directs the Service to preserve the natural 
resources, processes, systems, and values of units of the national park system in an unimpaired, evolving 
condition. More specifically, they direct the Service to: 

• preserve the natural resources, processes, systems, and values of units of the national park system 
in an unimpaired condition, to perpetuate their inherent integrity and to provide present and future 
generations with the opportunity to enjoy them: 

• recognize that natural processes and species are evolving and allow this evolution to continue, 
minimally influenced by human actions; and 

• apply the tenn "natural condition" to mean the condition of resources that would occur in the 
absence of human dominance over the landscape. 



Natural Ecosystems and Native Species ^ 

The Service becomes more specific in its policy guidance for natural ecosystems and native species 
through four major concepts. The first concept provides a broad ecosystem overview : the Ser\ ice 
generally does not intervene in natural biological or physical processes - when it does, such as to remove 
human-impacts to natural ecosystem functioning, it bases its actions on clearly articulated, well-supported 
management objectives and the best scientific information available. The following specific statements 
guide the implementation of this ecosystem approach: 

• the Service will not intervene in natural biological or physical processes, with certain exceptions: 

• actions to restore natural ecosystem functioning that has been disrupted by past or ongoing human 
activities will use the minimum necessary interventions to achieve the stated management 
objectives; 

• biological or physical processes altered in the past by human activities may need to be actively 
managed to restore them to a natural condition or to maintain the closest appro.ximation of the 
natural condition in situations in which a truly natural system is no longer attainable: 

• landscape and vegetation conditions altered by human activity may be manipulated where the 
park management plan provides for restoring the lands to a natural condition; 

• revegetation efforts usually will use propagules representing species and gene pools native to the 
ecological portion of the park in which the restoration project is occurring but may use improved 
varieties or closely related native species for a natural area so degraded that restoration w ith gene 
pools native to the park has proven unsuccessful: 

• because naturally ignited fire is a natural process in many ecosystems sustained in parks, each 
park with vegetation capable of burning will prepare a fire management plan addressing natural 
and cultural resource objectives; safety considerations for park visitors, employees, neighbors, 
and developed facilities; and potential impacts to public and private property adjacent to the park; 
and 

• the extent and degree of management actions taken to protect or restore park ecosystems or their 
components will be based on clearly articulated, well-supported management objectives and the 
best scientific information available. 

The second major concept addresses species population dynamics: the Serv ice generalK does not 
intervene in native plant and animal species dynamics and natural fluctuations in their populations. 
Components of this concept include: 

• native species are species that have occurred or now occur as a result of natural processes on 
lands designated as units of the national park system and so native species in a place are evolving 
in concert with each other: 

• natural processes are relied upon to maintain native plant and animal species and to inHuence 
natural fluctuations in populations of these species: 

• the Service maintains as parts of park natural ecosystems all native plants and animals by: 

o preserving and restoring natural abundances, diversities, dynamics, distributions, habitats, 

and behaviors of native plant and animal populations and the communities and 

ecosystems in which they occur: 
o restoring native plant and animal populations in parks when thc\ ha\e been extirpated by 

past human-caused actions; and 
o minimizing human impacts on native plants, animals, populations, communities, and 

ecosystems, and the processes that sustain them. 



The third major concept addresses protection of the full genotypic range of native plant and animal 
populations in the parks. Features of this concept include: 

• individual plants and animals found within parks are genetically parts of species populations that 
^ may extend across both park and non-park lands; providing for the persistence of a species in a 

park may require maintaining a number of local populations, often both within and outside the 
park; 

• protecting the full range of genetic types (genotypes) of native plant and animal populations in 
the parks is achieved by perpetuating natural evolutionary processes and minimizing human 
interference with evolving genetic diversity; 

• steps for protecting species native to national park system units that are listed under the 
Endangered Species Act include: 

o active management programs to inventory, monitor, restore, and maintain habitats of 
listed species, control detrimental non-native species, control detrimental visitor access, 
and re-establish extirpated populations as necessary to maintain the species and the 
habitats upon which they depend; 

o cooperation with other agencies, states, and private entities to promote candidate 
conservation agreements aimed at precluding the need to list species; and 

o developing management actions for the protection and perpetuation of federally, state, or 
locally listed species through the park management planning process, including 
consultation with lead federal and state agencies as appropriate; 

• intervention to manage individuals or populations of native species only when: 

o such intervention will not cause unacceptable impacts to the populations of the species or 

to other components and processes of the ecosystems that support them; 
o management is necessary to protect rare, threatened, or endangered species; or 
o removal of individuals or parts thereof is part of an approved research project; is done to 
provide propagules for restoring native populations in parks or cooperating areas without 
diminishing the viability of the park populations from which the individuals are taken; or 
meets specific park management objectives; 

• restoration of extirpated native plant and animal species to parks whenever all of the following 
criteria are met: 

o adequate habitat to support the species exists or can be restored in the park, and if 

necessary also on adjacent public lands and waters, and, once a natural population level is 

achieved, the population can be self-perpetuating; 
o the species does not, based on an effective management plan, pose a serious threat to the 

safety of people in parks, park resources, or persons or property outside park boundaries; 
o the genetic type used in restoration most nearly approximates the extirpated genetic type; 

and 
o the species disappeared, or was substantially diminished, as a direct or indirect result of 

human-induced change to the species population or to the ecosystem; 

• the need to maintain appropriate levels of genetic diversity will guide decisions on what actions 
to take to manage isolated populations of species or to enhance population recovery; and 

• actions to transplant organisms for purposes of restoring genetic variability through gene flow 
between native breeding populations will be preceded by an assessment of the genetic 
compatibility of the populations. 



The fourth major concept addresses methods for obtaining propagules for restoring plant species to parks: 

• programs to restore plant species may incUide propagating plants in greenhouses, gardens, or 
other confined areas to develop propagules for restoration efforts or to manage a population's 
gene pool. 



Pest Management 

The Service provides specific policy guidance regarding pest management. It relies on integrated pest 
management (IPM - a decision-making process that coordinates knowledge of pest biology, the 
environment, and available technology to prevent unacceptable levels of pest damage, by cost-effective - 
means, while posing the least possible risk to people, resources, and the environment) to guide managing 
pests in parks. It monitors use of pesticides (any substance or mixture that is used in any manner to 
destroy, repel, or control the growth of any viral, microbial, plant, or animal pest) in parks through case- 
by-case review of pesticide use requests, taking into account environmental effects, cost and staffing, and 
other relevant considerations. It allows use of a chemical, biological, or bio-engineered pesticide in a 
management strategy following a determination by a designated IPM specialist that such use is necessary, 
and that all other available options are either not acceptable or not feasible. 

Managing Non-Native Species 

The Service identifies as exotic (non-native, alien, or invasive) species those species that occupy or could 
occupy park lands directly or indirectly as the result of deliberate or accidental human activities. The 
Service devotes significant management attention to exotic species because these are species that did not 
evolve in concert with the species native to the place, are not a natural component of the natural 
ecosystem at that place, and, as a result, threaten the naturalness of the ecosystem being preserved to the 
degree that they out-compete the native species or alter the natural processes of the ecosystem. 

The Service's goal for managing exotic species is to not allow them to displace native species if 
displacement can be prevented. In general, new exotic species will not be introduced into natural 
ecosystems in parks while, in rare situations, an exotic species may be introduced or maintained to meet 
specific, identified management needs when all feasible and prudent measures to minimize the risk of 
harm have been taken. Such deliberate introductions may occur when the species is: 

• a closely related race, subspecies, or hybrid of an extirpated native species; or 

• an improved variety of a native species in situations in which the natural variety cannot survive 
current, human-altered environmental conditions: or 

• used to control another, already- established exotic species. 

In some situations, exotic plant and animal species are maintained to meet an identified park purpose. In 
all other situations, exotic plant and animal species that do not meet an identified park purpose will be 
managed — up to and including eradication — if (I ) control is prudent and feasible, and (2) the exotic 
species: 

• interferes with natural processes and the perpetuation of natural features, native species or natural 
habitats: or 

• disrupts the genetic integrity of native species: or 

• disrupts the accurate presentation of a cultural landscape: or 

• damages cultural resources: or 

• significantly hampers the management of park or adjacent lands. 



For species requiring management, high priority will be given to managing those exotic species that have, 
or potentially could have, a substantial impact on park resources, and that can reasonably be expected to 
be successfully controlled. Lower priority will be given to exotic species that have almost no impact on 
park resources or that probably cannot be successfully controlled. 

The decision to initiate management is based on a determination that the species is exotic. For species 
determined to be exotic and where management appears to be feasible and effective, parks evaluate the 
species" current or potential impact on park resources; develop and implement exotic species management 
plans according to established planning procedures; consult, as appropriate, with federal and state 
agencies; and invite public review and comment, where appropriate. In designing programs to manage 
exotic species, parks seek to avoid causing significant damage to native species, natural ecological 
communities, natural ecological processes, cultural resources, and human health and safety. 

Soil Management 

Any program to restore plants to natural systems must recognize and provide for soil management to the 
degree that the natural soil condition has been disrupted by past human activities. As a result, the Service 
seeks to understand and preserve the soil resources of parks, and to prevent, to the extent possible, the 
unnatural erosion, physical removal, or contamination of the soil, or its contamination of other resources. 
Where necessary, the Service may import off-site soil or soil amendments to restore damaged sites where 
such use of a soil, fertilizer, or other soil amendment may be appropriate, provided that the use does not 
unacceptably alter the physical, chemical, or biological characteristics of the soil, biological community, 
or surface or ground waters. Soil obtained from off-site normally will be salvaged soil, not soil removed 
from pristine sites, unless the use of pristine site soil can be achieved without causing any overall 
ecosystem impairment to the donor site. 



PHILOSOPHICAL FOUNDATION 

The conceptual goal of seeking to manage parks to achieve the natural condition as defined in the 
Management Policies clearly is impractical to achieve given the already existing degree of human 
dominance over the entire earth. However, identitying the natural condition as the desired condition is 
useful. With respect to the question of restoring the chestnut to parks, such identification gives park 
managers a clearly stated, measurable goal towards which to direct their scientific study and resource 
management efforts. Park managers address this goal by developing achievable intermediate goals and 
practical steps for achieving those intermediate goals. The remainder of this paper focuses on practical, 
policy-appropriate science and management steps that warrant consideration in efforts to develop a 
chestnut restoration plan and supporting program. 



APPLICATION OF NPS MANAGEMENT POLICIES TO RESTORATION OF CHESTNUT TO 

PARKS 

Role of Science 

Decisions about natural resource management are based on planning supported by scientific and scholarly 
infomiation, environmental evaluation, and public involvement. Scientific activities of inventory, 
monitoring, research, and assessment are important components of a chestnut restoration program because 
they: 



• contribute to developing a long-range strategy; __, 

• guide the functioning of interdisciplinary teams and processes; 

• permit articulating the desired future conditions for the park's natural resources; ^_^- 

• provide the tools for obtaining and integrating the best available science; 

• generate understanding of the effects of management actions on natural resources whose function 
and significance are not clearly understood; ^ 

• provide the framework for applying long-term research or monitoring in an adaptive management 
context to evaluate results; 

• provide the data for fully and openly evaluating environmental costs and benefits and, through 
public involvement, incorporating mitigation measures; and 

• underlie planning for clearly avoiding impairment of park natural resources and values. -^ 

Potential Role of Special Designation Areas 

The Management Policies make available to parks two special site designations (Research Natural Areas 
and Experimental Research Areas) that could be used to facilitate and focus efforts to restore the chestnut 
to parks. Research Natural Areas are sites within parks that contain prime examples of natural resources 
and processes, including significant genetic resources, and that have value for long-term observational 
studies or as control areas for manipulative research taking place outside the parks. Experimental 
Research Areas are specific tracts in limited situations that are managed for approved manipulative 
research, which is research involving conscious alteration of existing conditions as part of the 
experimental design. 

Activities in Research Natural Areas generally will be restricted to non-manipulative research, education, 
and other activities that will not detract from an area's research values. Activities in Experimental 
Research Areas involve a greater degree of manipulation as part of the research design but also can 
include other potential uses, such as education or other activities that will not detract from the area's 
research purpose. 

Decisions and Actions Involve Partners 

NPS fully recognizes that many organizations are involved in efforts to restore the chestnut to the forests 
of the United States. Management Policies encourage park managers to develop agreements 
appropriately with others to coordinate chestnut restoration activities in ways that would maintain and 
protect, not compromise, park resources and values, including the integrity of native gene pools and 
natural ranges of species and ecological communities. In entering into such agreements, park managers 
would be encouraged to work with other land managers to encourage the conservation of populations and 
habitats of the chestnut wherever and whenever possible, including through such NPS actions as: 

• participating in local and regional scientific and planning efforts, identify ing chestnut local 
population characteristics and ranges, and developing cooperative strategies for maintaining or 
restoring park components of these local populations; 

• preventing the introduction of new exotic species into units of the National Park System while 
removing populations of the chestnut blight that have alreadv become established in parks, and 

• providing small quantities of chestnut genetic material from parks for cooperators to use in 
selective breeding, genetic engineering, or propagule generation efforts. 

At the same time, the Policies encourage managers to avoid the dissemination into the wild of chestnut 
genetic material outside the native range of the chestnut, unless such dissemination is conducted under a 



specific, scientifically-based management program designed to mitigate for a human-facilitated 
environmental impact, such as habitat fragmentation or global climate change. 

Decisions and Actions to Restore the Chestnut 

Park actions to restore the chestnut would respond to clear goals, implement a proactive strategy, and be 
based on the clear responsibility park managers have to preserve natural conditions. Goals could include: 

• re-establishing in human-disturbed components of park natural systems (those where introduction 
of the chestnut blight has nearly eliminated a dominant native species) the natural functions and 
processes provided by the chestnut by restoring appropriate chestnut genotypes or the best 
available surrogates; 

• using the best available technology, within available resources, to restore the chestnut and, as a 
result, to stimulate restoration of its associated biological and physical system components and 
accelerate recovery of landscape and biological-community structure and function; and 

• removing the exotic species or at least greatly reducing its role in the ecosystem. 

Elements of an appropriate strategy would include: 

• maintaining existing, in-park, local populations of native genotype plants that continue to resprout 
following blight-induced death of their previous sprouts for the purpose of maintaining living 
genetic material for future research efforts 

• restoring the native species using organisms taken from populations as closely related genetically 
and ecologically as possible to park populations, preferably from similar habitats in adjacent or 
local areas, where possible; 

• introducing different native genotypes where the management goal is to increase the variability of 
the park gene pool to mitigate past, human-induced loss of genetic variability; guided by 
knowledge of local adaptations, ranges, and habitat requirements, and detailed knowledge of site 
ecological histories; 

• introducing novel, non-native genotypes where the management goal is to develop a gene pool 
that is genetically resistant to the chestnut blight, guided by a goal of inserting the resistance with 
as minimal an insertion of other non-native genes as possible; 

• applying the Service's integrated pest managemen; (IPM) program to eradicate, or at least control, 
the chestnut blight to whatever degree possible while reducing risks to the public, park resources, 
and the environment from chestnut blight and blight management strategies; and 

• utilizing appropriate soil conservation and soil amendment practices to facilitate restoring the 
chestnut in ways that prevent or minimize adverse, potentially irreversible impacts on soils. 

Given that the park is the basis and focus of NFS natural resource management programs, it is important 
to recognize that: 



'&' 



resource management is a local activity and the park superintendent exercises the responsibility 
for, and is held accountable for, all actions that occur within the park - therefore, cooperative 
actions to develop an NFS chestnut restoration activity would use the park and its partners as the 
basic building block; and 

the Service's use of networks of parks for inventory and monitoring purposes would offer a 
strategic opportunity for cooperatively applying metapopulation and biological corridor concepts 
to chestnut restoration efforts. 



DISCUSSION 

From this review of NPS Management Policies, it becomes clear that NFS polic\ is not an issue for 
determining whether or not to restore the chestnut to parks. Our current state of knowledge about the 
status of the chestnut meets key policy provisions: 



• extirpation is occurring and its cause is known to be an exotic species, hence the extirpation is 
human-caused and management restoration is appropriate: 

• the impact of the extirpation on park natural resources is apparent - loss of a dominant species, 
probability of cascading ecological effects, associated human social and economic effects, all of 
which possibly may constitute impairment; and -=^ 

• a management response is clearly possible - minimize the effect of the exotic species both by 
controlling the exotic species and by developing and planting seeds, seedlings, and saplings of a 
blight-resistant chestnut genotype. 

These policy provisions suggest a clear goal - restore a naturally functioning, natural ecosystem by 
restoring a nearly extirpated native species, eliminating the exotic species or at least neutralizing its 
impact on that native species, and restoring the ecosystem function once provided by the nati\ e species. 

The policy goal of maintaining a practically appropriate level of genetic ridelit\ can be met. First, 
although many of the original local population gene pools are so diminished they probably can not be 
restored, there are a few existing, endemic, apparently disease-resistant North American genotvpes that 
can be propagated and disseminated as a means of maintaining at least some native genetic material in the 
gene pools used for restoration. Second, specific, appropriate genes from several nearest-relative gene 
pools can be injected into the residually available North American gene pools either through cross 
breeding of North American and Asian genotypes or through using genetic engineering to insert selected 
foreign genes into the residual native genot> pes. 

If a decision were to be made to implement this policy of restoring a species and its associated ecosystem, 
there are clear science needs that must be met as part of planning. NEPA analysis, and developing 
restoration methods. These infonnation needs include: 

• understanding how the existing ecosystems and their current floras, faunas, and physical features 
might change if restoration were successful; 

• assessing whether any native species would become at risk if restoration were successful; 

• determining if there would be any risk of introducing other pathogens in association with planting 
cultivated seeds or young trees; 

• addressing what side effects, if any. individual park ecosystems or their local chestnut 
populations would experience as a result of addition of a genetically modified chestnut to the 
individual ecosystems; 

• determining whether any physical alteration of existing ecosystems would be needed to achieve 
an effective restoration and, if so, what would those alterations be; and 

• developing park chestnut restoration activities as scientific experiments with good design, 
methods, replication, and documentation - in essence, structuring these activities as adapli\e 
management. 

Carrying out a chestnut restoration program clearly would have to involve the public. The program 
would be a long term activity requiring support over many years. It would depend on the involvement 
and good will of man\ partners. It would have to be based on a clear understanding b\ all participants of 
the scientific basis for. and methodological requirements o\\ each of the possible management approaches. 



10 



Because of its ultimate wide spread distribution over the landscape and through time, its probability of 
success would depend on the level of stakeholder agreement, support, and participation maintained across 
space and through time. 

The NPS Management Policies provide a framework for determining what kinds of restoration 
management action might be appropriate, for ensuring that scientific findings play a significant role in 
informing the detemiination, and for broadly and effectively involving the public in the decision-making 
process. For whatever management program might be adopted, the Policies leave to the discretion of the 
site manager what specific mix of technologies to apply, with the mix at any given site being influenced 
by such site-specific factors as what the science shows to be technically possible, what the environmental 
analysis shows to be the trade-offs between environmental and human benefit and detrimental impact, 
what actions the public involvement reveals to be locally and generally acceptable, and what fiscal and 
human resources are likely to be available for conducting the management program over the projected 
duration of the restoration effort. 



CONCLUSION 

National Park Service management policies encourage restoration of the chestnut to park ecosystems. 
These policies require that such restoration be accomplished using a process that includes science, 
planning, and public involvement. These policies strongly encourage adopting a management program 
that emphasizes cooperation and collaboration with partners. 



BIBLIOGRAPHY 
National Park Service. 2000. Management policies 2001. National Park Service, Washington, DC. 137 p. 

ACKNOWLEDGEMENT 

I thank Greg Eckert for providing the workshop attendees with a condensed summary of these policies in 
my absence. 



11 



12 



Steiner, K. C. and Carlson, J. E, eds. 2006. Restoration of American Chestnut To Forest Lands - Proceedings of a 
--- Conference and Workshop. May 4-6. 2004, The North Carolina Arboretum. Natural Resources Report 
NPS/NCR/CUE/NRR - 2006/001. National Park Service. Washington. DC. 

J HISTORICAL ECOLOGY OF AMERICAN CHESTNUT {CASTANEA DENTATA) 

Emily W. B. (Russell) Southgate 
: v^ . Notre Dame Academy, Middleburg, VA 201 18 USA (erussell@nac.net) 

Abstract: American chestnut {Castcmea dentata) was a very common species in the forests of eastern 
North America in the early 20"' century when it was decimated by the introduced chestnut blight. The 
post-glacial migration history of chestnut differed from its most common associates, oak and hickory due 
most likely to differences in the ecological tolerances of the species. By 1 500, both pollen evidence and 
historical documents indicate that chestnut trees formed 5-15% of oak-dominated forests throughout the 
northeast, as far north as southern New England. The vigorous sprouting of chestnut after it is cut 
allowed it to develop widespread "sprout forests," where chestnut trees were 50% or more of the stems in 
many stands, after 18"^- 1 9"^ century logging. This high concentration of chestnut stems may have 
allowed the blight to spread very quickly throughout its range. In addition, many other changes have 
occurred in regional forests over the last century, as they have responded to a variety of human-caused 
disturbances. Thus, introduction attempts should take into account the time period that is of interest to try 
to restore and the dynamics of current forests in considering what might be the fate of chestnut trees that 
they reintroduce into today's forests. 

Key Words: historical ecology / sprout forests / pollen / historical documents 

INTRODUCTION 

Doomed sprouts of American chestnut (Castanea dentata) are widely dispersed throughout the hardwood 
and hemlock/hardwood forests of the eastern United States today. In 1900, mature chestnut trees 
dominated many of these stands, and were very common in many others. A species valued not only for 
its majestic beauty, but also for timber and fruit, chestnut trees were planted well beyond their natural 
range. A deadly chestnut blight from an introduced tree in New York City, however, destroyed all mature 
trees in the early 20"' century, leaving only ghostly stumps and shrubby sprouts as legacies of this once 
majestic tree. 

A brighter future for the American chestnut may now be in the hands of foresters who have developed 
strains resistant to the blight. In considering the possibility of restoring this native tree to its former 
habitats, it is important to consider its former range, and its unique history as it developed to its 
distribution at the time the blight killed the trees. There are two major reasons to consider this history. 

>- First, we need to establish an appropriate goal for restoration. Is there a specific period in 

the history of this species that is particularly desirable? History can provide analogues that may 
be considered as possible goals of restoration. This gives the restoration ecologists clear 
endpoints to consider. 

> Second, history can provide a picture of the changing distribution of the species, and of 

other species that were associated with it, over time, as they may both facilitate and constrain the 
likely outcome of restoration. Forests are always dynamic in their composition and structure, and 
understanding these dynamics can allow restoration ecologists to evaluate potential trends and the 
future of current changes. (Russell 1997). 

I will discuss the history of the changing range of American chestnut and the other species with which it 
was associated in two time periods: 1. from 10,000 BP to 1500 AD, as it migrated to its pre-Columbian 



n 



range; 2. from 1 500 AD to 1900 AD, after European settlement of the area. This discussion will be 
based on the record left by trees in the form of pollen preserved in lake sediments and on the historical ~ \ 
documentary record. My analysis will focus mainly on the area from northern New Jersey to northern ^^ - 
New England, as that is where the most records of the historical distribution have been compiled. 



METHODS 

Pollen 

Pollen preserved in lake sediments provides a unique record of the history of plants, especially in areas -_ 
which were covered with glaciers in the most recent (Wisconsinan) ice age. where disrupted drainage has 
left many sedimentary basins which have accumulated sediments over the millennia. Many tree genera, 
such as pines {Finns) and oaks (Quenus), are pollinated by wind, so produce copious amounts of airborne 
pollen. Others, such as maples (Acer), are pollinated by insects, and produce less pollen, since pollination 
is more assured when the insect carries pollen from tree to tree. American chestnut seems to fall in an 
intermediate category of pollen production and dispersal. While insects do visit the male catkins of 
chestnut trees, the trees produce large amounts of windblown pollen. 

Pollen is identifiable under the microscope to varying degrees of specificity'. Pollen analysts can 
distinguish oak and hickory pollen only to genus. Chestnut is also identifiable only to genus. However, 
only two species are found in eastern North America, and of these only Castanea dentata is widespread 
and exists as a large tree that would distribute pollen any distance from the tree that produced it (Paillet 

2000). 

After pollen is released by a tree, it is carried by the air before falling to the ground. Even wind- 
pollinated species drop a large proportion of their pollen within a few hundred meters of the tree. 
After pollen grains land on a body of water, they eventually sink to the bottom and are incorporated into 
the sediments. As sediments build up over the years, they thus contain a record of the trees and other 
plants that have grown in the vicinity as well as of the openness of the vegetation. Pollen grains are very 
resistant to decay in such as situation, and can provide a proxy for reconstructing past plant, especially 
tree, distributions. 

Large bodies of water, greater than a hectare or so, collect windblown pollen from large regions. lO's of 
kilometers from the lake, because there is a large ratio of surface area/shoreline. Smaller ponds and 
hollows, on the other hand, may collect pollen from mainly local sources, 50 m or so from the 
sedimentation site. These differences allow us to reconstruct vegetation that has produced the pollen on 
both a regional and a very local, stand-level, basis. 

When pollen is produced before leaves, as in the oaks, the wind currents often carr\ the pollen many 
kilometers. Chestnut trees release their pollen after the leaves ha\e expanded, which means that the 
pollen, though copious, is often caught by leaves, and does not enter the air cuirents. This allows us to 
interpret chestnut distribution from pollen preserved in sediments on a finer scale that we can detemiine 
for many other species which produce large amounts of pollen. Finally, pollen of plants that grow close 
to the ground in a forested landscape is not carried far. and mostK falls directly to the ground. If the trees 
are cut, however, pollen produced below a meter b\ plants such as grasses (Poaceae) and ragweed 
{Amhrosia) can be carried several kilometers. The recent ecological history of eastern North America, 
characterized by massive regional deforestation after the arrival of European settlers can be dated b> 
increases in these weedy species, even when looking at areas that were not locally deforested. 



14 



The pollen data that I will discuss come from three sources: 

• First, for the brief discussion of the millennial record. I will use data compiled by Thompson Webb 
__^- 111, and published in numerous publications for interpreting many aspects of post-glacial vegetation 
::^ and climate trends. Specific references to the pollen collections can be found in Bernabo and Webb 

(1977). 

• The second set of data are those compiled by Russell and Davis (Russell et al 1993, Russell and 
Davis 200 1 ), which include more detailed records of species distributions over the last 500 years, 
focusing on human impact on species distributions. These data only cover the area from northern 
New Jersey north to north-central Maine - the area covered by Wisconsinan glaciation. The trends 
detected in these data are, however, most likely similar to trends farther south, though further study 
may either confinn or refute that speculation. 

• Finally, 1 will discuss data from some very small sites in central Massachusetts, which allow the 
reconstruction of very detailed stand histories (Foster and Aber 2004). 

Historical Documents 

After the arrival of European settlers in North America, written documents serve also to trace the history 
of the distribution of forest species. Early travelers provided qualitative descriptions of forest resources, 
often with excellent taxonomic accuracy. These almost always, however, lack any quantitative 
information. The earliest quantitative data come from land surveys, generally what are referred to as 
"metes and bounds" surveys which delineate properties. Surveyors were trained to recognize local tree 
species, and often used trees as markers for property boundary comers. The parts of the United States 
settled after the American Revolution were surveyed according to a very clearly codified rectilinear 
survey, but the colonial lands were surveyed by a variety of methods, some quite systematic and others 
much less so. By assembling these data, we can, however, obtain a remarkably consistent record of the 
species distributions in the precolonial forests, before settlers cleared them for farms (Loeb 1987, 
Whitney 1994. Russell 1997, Cogbill 2000). 

The second set of documents that can provide evidence for the preblight distribution of American 
chestnut is the plethora of forest surveys around the end of the 19"' century by state surveyors. The states 
had begun to realize that careless logging, grazing and fires had severely damaged their forest resources. 
To evaluate the problem, they embarked on systematic surveys to provide infomiation that could guide 
their efforts to protect and improve their forests. These provide a snapshot of the condition, composition 
and structure of the forests of this period, when most heavy logging had moved away from the original 1 3 
colonies, leaving regenerating, generally young forests in the east, especially the northeast (Russell 1987). 



DISCUSSION 

From the end of the Wisconsinan glaciation to 1 500 AD 

When the extreme cold of the Wisconsinan glaciation dominated the northern half of North America, tree 
species that today characterize forests north of the terminal moraine ranged far to the south, where 
climates were considerably colder than they are today. They were found in novel assemblages, depending 
on the local climate and the ecological tolerances of species for these conditions (Webb 1988. Delcourt 
and Delcourt 1987). The sketchy pollen record from this time period indicates that American chestnut 
was a fairly minor component of forests dominated by oak, along with some other associates such as 
hemlock {Tsiiga canadensis)ov black gum {Nyssa sylvatica) (Barclay 1957, Bender et al. 1979, Craig 



15 



1969) in the southeastern Appalachian region. In Horse Cove Bog, western North Carohna, however, it 
was represented in quantities of pollen almost equal to oak between about 1400-150 BP (H.R.Delcourt 
and P.A.Delcourt, Pers. Comm, University of Tennessee. 1996). As climate moderated, the range of ^ 
chestnut expanded slowly northeastward along the Appalachian and Ridge and Valie\ Provinces. ~~" 
reaching very large concentrations in some places before again declining (Barclay 1957, Davis 1983, 
Webb 1988). -^_^..-^ 

The spread of chestnut into the northern forests lagged behind the oaks and hickor>' {Caryd). For 
example, in southern New York oak had reached its current importance in the pollen record about 9000 
YBP, while chestnut did not appear above 1% or so until about 4000 YBP (Maenza-Gmelch 1997). 
Likely explanations include different climate tolerances coupled with its more demanding pollination 
mechanism. Because American chestnut cannot self pollinate, a single tree growing beyond the current 
range of the species could not produce fertile seeds to spread from this point, while a hickorv or oak tree 
could do so. 

It is also possible that this distinct history is an artifact of studying all species of oak at one time, because 
they cannot be distinguished in the pollen record. The different species of oak represented in the east 
have quite varied ecological tolerances, while we can assume that we are tracing just one species in the 
case of chestnut. After about 2500-2000 BP, chestnut reached its current range. There is some recent 
evidence based on lake levels correlated with pollen data that it spread north as climate became more 
humid after 2000 BP (Shuman et al. 2004). 

1500 AD to the present 

By 1 500 AD. chestnut was a consistent member of the oak-dominated forests of many eastern forests, 
according to the pollen record (Russell et al 1993, Davis 1983, Webb 1988). It has a much more 
restricted range than oak or hickory throughout its postglacial historv, being restricted to a fairly narrow 
band along the Appalachian physiographic province (Davis 1983). Again, this may in part be due to 
comparing all the species in one genus to one species. By 1 500 AD, chestnut contributed 4-9% of the 
pollen in lake sediments south of northern Massachusetts, where oak and hickoiy were the dominant taxa 
in the forests. North of about 43"N, where spruce (Picea), pine, birch (Bctiila). hemlock and beech 
{FagKs grandifo/la) dominated the forests, chestnut was generally less than one percent of the pollen 
indicating that it was at most a minor component of the forests (Russell and Davis 2001 ). 
Historical records confirm and expand upon these pollen data. According to data compiled by G. Gordon 
Whitney, American chestnut trees were generally 5-15% of the trees listed in early land sur\e\s in 
Pennsylvania, eastern Ohio, northern New Jersey, extreme southeastern New York. Long Island. 
Connecticut, Rhode Island and the Connecticut River valley in Massachusetts. White oak {Qucrcus alba) 
dominated these forests, with 25-65%) of the stems, along with 5-15% hickor). These data have not yet 
been compiled from farther south in the range of the species. 

A breakdown of the data from the area from northern New Jersey and to western Massachusetts shows 
some details of this distribution (Table I ). Chestnut was most common in forests dominated by oak. with 
little hemlock or beech, while it was less common (though occasionally present) in areas where beech and 
hemlock dominated the forests. In eastern West Virginia, chestnut was most common on ridgetops, 
where it formed 15% of witness trees, compared with 2-5% in other topographic positions (Abrams and 
McCay 1996). In Pennsylvania, chestnut was most common in the Allegheny Mountain ph\siographic 
province, though present throughout the state. Here, too, it was most common on hilltops (Abrams and 
Ruifner 1995). 



16 



Table 1. Percent of trees in precolonial kind surveys in northern New Jersey, eastern Pennsylvania, 
eastern New York arid western Massachusetts (data from Russell 1981, Biirgi et al. 2000, Mcintosh 1962 
and unpublished data for the Shawangunk Mts. and Rensselaerville, NY) 





n.e. NJ 

(Morris 

Co.) 


n.e. PA 
(Pike 
Co.) 


n.e. PA 

(Wayne 

Co.) 


e.NY 

(Shawangunk 

Mts.) 


e.NY 
(Catskill 

Mts.) 


e.NY 
(Rensselaerville) 


w. MA 

(Berkshire 

Co.) 


Chestnut 


15 


7 


1 


7 








6 


Oak 


64 


40 


6 


41 





4 


16 


Beech 


1 


3 


36 


1 


50 


48 


23 


Hemlock 





5 


22 


4 


20 


14 


19 


Maple 


4 


7 


16 


6 


14 


14 


11 


Pine 





27 


3 


6 





1 


7 


Total 
number 
of trees 


199 


1921 


939 


342 


3744 


114 


1730 



Some details of local distributions and response to disturbances have been found in studies in central 
Massachusetts (Foster and Aber 2004). In these studies, pollen from small hollows or mor humus soils 
reflects the proportion of trees growing within 50 meters or so of the sampling point. Chestnut 
importance appears to have alternated with oak where oak was dominant. In another site, it appears that 
chestnut responded quickly to disturbance, but was supplanted by hemlock after the chestnut blight. 

Whatever further study may reveal, however, it appears that on a broad scale, of a county or more, 
chestnut was a consistent but fairly minor associate of oak, especially white oak in the precolonial forests. 
How can we reconcile this with evidence of forests dominated by chestnut at the turn of the 19"' century 
(Russell 1987)? The answer lies in the physiology of the species, in particular, its tendency to sprout 
vigorously from the root crown when it is cut (Paillet 2000). 

Between the first settlement of the eastern United States by European settlers and 1900, the new 
inhabitants cleared all but the most remote and difficult to reach forests. Some land was turned into 
farms, but much that was not good agricultural land was repeatedly cut over for fuel and timber, 
especially for making charcoal to feed iron forges and furnaces. The forests of the first half of the 20"' 
century were designated by E. Lucy Braun as "sprout hardwoods" referring to this tendency to sprout 
(Braun 1950). There is evidence in the pollen record for this change in the importance of chestnut in the 
forests of the northeast (Russell et al 1993, Russell and Davis 2001 ). Chestnut is one of the species that 
consistently increases in proportion of the tree pollen after the increase in agricultural indicators in the 
pollen record. It is not a major pollen producer like birch, which also increased, so the apparently small 
increase recorded in the pollen most likely translates into a much greater increase in the proportion of 
trees in the forest. 

It is likely, therefore, that the forests that the blight decimated were primed to spread a pest like this. 
While not fonning a monoculture, the species was very common by this time, and thus allowed the blight 
to spread quickly throughout its range (Russell et al. 1993). The distribution and abundance of sprouts in 
forests today reflect a forest greatly modified by the impact of European settlers. That these sprouts 
represent seedling trees, not the forest giants, is even more suggestive of the dynamic position of chestnut 
in these early 20"^ century forests (Paillet 2000). 



17 



Today's forests reflect this complex history. Chestnut sprouts abound, and their distribution indicates the 
sites that are most appropriate for chestnut to succeed. Disturbance is a positive force for chestnut 
growth. The current forests of the United States have changed significantiv in species composition in the 
last 500 years, with a general decrease in the amounts of hemlock and beech and an increase in birch. 
Given the associates of chestnut in the historical record and its responses to disturbance, it seems likely 
that it would respond well to current conditions. _-, - 



LITERATURE CITED 

Abrams. M.D., and CM. Ruffner. 1995. Physiographic analysis of witness-tree distribution (1765-1798) 
and present forest cover through north central Pennsylvania. Can. J. For. Res. 25:659-668. 

Abrams. M.D., and D.M. McCay. 1996. Vegetation-site relationships of witness trees (1780-1856) in the 
presettlement forests of eastern West Virginia. Can. J. For. Res. 26:217-224. 

Barclay, F.H. 1957. The natural vegetation of Johnson county, Tennessee, past and present. Ph.D. thesis, 
The University of Tennessee, Knoxville. TIM. 

Bender, M.M., D.A. Baerrels, and R.A. Bryson. 1979. University of Wisconsin Radiocarbon Dates XVI. 
Radiocarbon 21:120-130. 

Bernabo. J.C.. and T. Webb III. 1977. Changing patterns in the Holocene pollen record from northeastern 
North America: a mapped summary. Quatemar> Res. 8:64-96. 

Braun, E.L. 1950. Deciduous forests of eastern North America. The Free Press, New York. 

Biirgi. M., E. W. B. Russell, and G. Motzkin. 2000. Effects of postsettlement human activities on forest 
composition in the north-eastern United States: a comparative approach. J. Biogeography 27:1 123-1 138. 

Cogbill, C.V. 2000. Vegetation of the presettlement forests of northern New England and New York. 
Rhodora 102:250-276. 

Craig, A.J. 1969. Vegetational History of the Shenandoah Valley, VA. GSA Special Paper 123:283-296. 

Davis, M.B. 1983. Holocene vegetational historv' of the eastern United States. P. 166-181 in. Late 
Quaternary Environments of the United States. Vol. 2., Wright, H.E., Jr. (ed.). The Holocene. University 
of Minnesota Press. Minneapolis. MN. 

Delcourt, P. A., and H.R. Delcourt. 1987. Long-term forest dynamics of the temperate zone. Ecol. Studies 
63, Springer- Verlag, New York, NY. 

Foster, D.R., and J.D. Aber. 2004. Forests in time. The environmental consequences of 1.000 years of 
change in New England. Yale University Press. New Haven. CT. 477 p. 

Loeb, R.E. 1987. Pre-European settlement forest composition in east New Jersey and southeastern New 
York. Am. Midi. Nat. 118:414-423. 

Maenza-Gmelch. T.E. 1997. Holocene vegetation, climate, and Ine hislorv of the Hudson highlands, 
southeastern New York, USA. Hie Holocene 7:25-37. 



18 



Mcintosh, R.P. 1962. The forest cover of the Catskill mountain region. New York, as indicated by land 
survey records. Am. Midi. Nat. 68:409-423. 

Paillet, F.L. 2000. Chestnut: history and ecology of a transformed species. J. Biogeography 29: 1 5 1 7- 
1530. 

Russell. E.W.B. 1981 . Vegetation of northern New Jersey before European settlement. Am. Midi. Nat. 
105:1-12. 

Russell, E.W.B. 1987. Pre-blight distribution of Castanea dentata (Marsh.) Borkh. Bull. Torrey Bot. Club 
114:183-190. 

Russell. E.W.B. 1997. People and the land through time: linking ecology and history. Yale University 
Press. 306 p. 

Russell, E.W.B., and R.B. Davis. 2001. Five centuries of changing forest vegetation in the northeastern 
United States. Plant Ecol. 155:1-13. 

Russell, E.W.B.. R.B. Davis. R.S. Anderson, T.E. Rhodes, and D.S. Anderson. 1993. Recent centuries of 
vegetational change in the glaciated northeastern United States. J. Ecol. 81:647-664. 

Shuman. B., P. Newby, Y. Huang, and T. Webb 111. 2004. Evidence for the close climatic control of New 
England vegetation history. Ecology 85: 1 297- 1310. 

Webb, T. III. 1988. Eastern North America. P. 385-414 in Vegetation History, Huntley, B., and T. Webb 
III (eds.). Kluwer Academic. 

Whitney, G.G. 1994. From coastal wilderness to fruited plain. A history of environmental change in 
temperate North America 1 500 to the present. Cambridge University Press, Cambridge. 45 1 p. 



20 



Steiner, K. C. and Carlson. J. E, eds. 2006. Restoration of American Chestnut To Forest Lands - Proceedings of a 
Conference and Workshop. May 4-6, 2004, The North Carohna Arboretum. Natural Resources Report 
NPS/NCR/CUE/NRR - 2006/001, National Park Service. Washington, DC. 

FOREST HEALTH IMPACTS OF THE LOSS OF AMERICAN CHESTNUT 

(transcript of presentation) 

x^ . Steve Oak 

USDA Forest Service, Forest Health Protection, 
P.O. Box 2680, Asheville, NC 28802 USA (soak@fs.fed.us) 



INTRODUCTION 

I consider thus meeting and these topics - the restoration of American chestnut and trying to return 
chestnut to an important ecosystem component in southern Appalachia or in the Appalachians of the 
Eastern United States - to be very important. My life has been consumed recently with sudden oak death. 
I leave tomorrow to train the last of three groups of people involved in an expanded early detection survey 
for sudden oak death. I open that training by helping people visualize what it might have been like in the 
days, weeks and months after the initial discovery of chestnut blight in the Bronx Zoo and it kind of 
brings the message home. We don't know if that is what is going to happen with sudden oak death, but it 
does represent one of the possible outcomes on a continuum from innocuous or no impact to a chestnut 
blight type of scenario. 



FOREST HEALTH 

"Health and integrity are not inherent properties of an ecosystem and are not supported 
by either empirical evidence or ecological theory." (Wicklum and Davies 1995) 

I will start with a bit of discussion on forest health. This would be a short talk if this was the definition 
that we accepted - that essentially there is no such thing as forest health. Health and integrity are not 
inherent properties of an ecosystem and are not supported by either empirical evidence or ecological 
theory. I could stop there, but I won't. 

A draft Forest Service Policy was set forth in 1996 and '97. It is still a draft because when you try to 
bring diverse groups of people together and come to consensus, you end up with chaos, usually. This is 
still a draft policy, but there are elements that I want you to think about as we walk through the talk. 

Forest health is measured at a landscape scale. We are not talking about tree health or even stand health, 
but forest health - consider it on a landscape scale. The notion of forest health carries with it the idea of 
ecological integrity and that forest components and relationships are all present, functioning, and self- 
renewing. And you can imagine what the elimination of chestnut as a functioning ecosystem component 
did in the early part of the 20"^ century - how was that affected, that ecological integrity component? 
Forest health also has a human dimension - the idea that forests should provide for human values, uses, 
products, and services. And those values etc. are fluid; they change with our ideas about why forests are 
important. 

It's appropriate that the previous presentation was about forest history' and what was originally here. Ten 
thousand years ago the forest composition was quite different from what it is today. That pushes back the 
perspective from what is was like at European settlement and when the first native people were in this 
area to the idea that these forests are nothing if not ever-changing as a result of the way that people 
interact with the forest. I will start with this supposition that southern Appalachian forest landscapes are 



21 



unprecedented in history. There's never been anything like what you see here today. And the forest that 
will result in decades hence from what is there today will be like nothing else that has ever existed in the 
past. The components of these ecosystems were already in place. Txe read, about 58 million years ago, 
ebbing and (lowing with ice sheets and fire. But it reall> wasn't until the last 1 0,000 \ ears or so. or 
maybe even more recently than that, that we've had forests that resemble in structure and composition 
what was present at European settlement. So why has there never been anything like what we see here 
today? Of course one important thing, and perhaps the most significant element, was the introduction of 
the chestnut blight, with ground zero at the Bron.x Zoo in 1904. 



PROGRESSION OF THE CHESTNUT BLIGHT 



It may be something of a fallacy to think of the chestnut blight moving through the eastern hardwood 

forests in a wave, nothing in front of this wave and 
devastation behind it. But 1 draw your attention to a little 
spot of infection in Bedford Count\\ Virginia (circle in 
Figure I ), four or five years after discovery of the blight in 
New York, well in front of the general advance. There is 
no way that occurred from a continuous spread, and 1 
suggest that this and many other infection sites were the 
result of subsequent introductions or movement of infected 
material either prior to or after the discovery in New York. 

In 1911, the infestation in Bedford Count). Virginia, was 
well ahead, or outside of. what might be referred to as the 
advancing front (Figure 2). Maps from literature 
published in the 1920"s about the progress of the blight 
through the southern Appalachian assessment area show 
infection in Greenville Countv. South Carolina/Henderson 
Count> North Carolina, in 1926 (Figure 3). It was known 
that at the border between Polk, which is the count> 
immediatel) to the East of Henderson and Greenville 
Count>. there was an infestation dated back to 1912. based 
on the regular increments and dating of cankers at that location. So, there in 1912. and in 1908 in Bedford 
County, shows that it was not a continuous spread, not an even wave running through the s\ stem. 




Figure I. Chestnut blight distribution in 
1909 (Metcalf and Collins 1909). = 
Bedford Countv. VA. 



The blight wasn't the only thing going on (in the woods) at that time. There was heavy dut> forest 
utilization. What I tr\ to point out to people, is that what fomis the structure of toda\"s forest is a result 
of not just the chestnut blight, because land use practices and events immediately prior to and just after 
the chestnut blight were very important as well. Some of those were fire and heavy utilization, and then 
with regard to fire, not just the presence of fire, but then the almost complete absence of fire following the 
Weeks Act and the formation of the National Forest and Cooperati\e Forest Fire Control Programs in the 
states. So you went from a heavy disturbance regime, introducing chestnut blight on top of that, and then 
ceasing most heavy disturbance activities. 

Table 1 summarizes the pre- 1900 and current conditions of the southern Appalachian forests. The 
southern Appalachian forests before 1900 were dominated by American chestnut in man\ places. 
Whether this was an artifact of disturbance by native people or earl\ European acti\ ities is less relevant 
than what was there and being impacted at the time. But ain where from a quarter to a third, depending 
on the inventory' that you read from the period, in this core Appalachian area of North Carolina- 
Tennessee-North Georgia-Virginia, had sparse understories, large, w ideh spaced overstories, and a high 



22 



level of disturbance from fanning, logging, and fire. When fire regimes were altered, and with oaks 
already an important part of the forest, oaks were positioned to take the newly available space that was 
made available with the loss of chestnuts. So now we have dense understories, dense overstories of 
somewhat smaller diameter trees, and relatively low disturbance regime as compared with the historical 
past. And then there was the introduction of the gypsy moth, a non-native defoliator, fires suppression 
programs and a growing human population. These are the backdrops against which we interpret forest 
health changes. 




Figure 2. Chestnut blight distribution in 1911 
(Metcalf 1912). 

Figure 3. Chestnut blight epidemic in the 
southern Appalachians in 1926. 




Table 1. Composition, structure, and disturbance profiles of southern Appalachian forests, pre 1900's vs. 
present day. 



Pre- 1900 


Current 


American chestnut 


Aging oak cohort 


Sparse understor> 


Dense understor>' 


Large overstory 


Dense overstor> 


High disturbance 


Low disturbance 


Farming 


Gypsy moth 


Logging 


Fire suppression 


Fire 


Human population 



23 



FOREST HEALTH IMPACTS OF THE LOSS OF AMERICAN CHESTNUT 

This is a quote from Smith (1976) from the "Changes in Eastern Forests" article in Perspectives in Forest 
Entonwlogy: 

"We are perhaps entitled to speculate that our chronic ami alarming problems with the 

gypsy moth ami other oak defoliators in the eastern or Appalachian portions of the mixed - — - -"' 

deciduous forest could be as evil a consequence of the chestnut blight as the loss of 

chestnut itself. " 

Oak dechne is a disease that I will be discussing as a major forest impact of chestnut blight. Again, these 
oaks came in as a relatively even-age cohort after the loss of chestnut. They have grown up pretty much 
without disturbance since. People who drive the parkway up on the ridge above you here look out at the 
landscape and think that it has always looked like this. It is wonderful that we have this preserved area, 
but in fact this landscape is probably less than 100 years in the making. 

What is oak decline? 

The symptoms of oak decline are a progressive dieback from the top down and outside in. on dominant 
and codominant oaks trees that have proved their competitive metal over the decades. Again, decline is 
progressive from the standpoint that it may take years or even decades to progress from those initial 
symptoms to more advanced symptoms. In late stage symptoms >ou have epicomiic sprouts coming off 
the main stem. There can be a gradation of twig condition, from twigs that still have buds on them, and 
are very recently dead, to branches that have dieback. But these are signs of a progressive dieback. taking 
years or even decades, progressing to mortality in susceptible trees. The species in the red oak group are 
more susceptible to oak decline mortality than those in the white oak group. 

According to Sinclair (1965), oak decline etiology begins with factors that predispose the tree to decline 
(predisposing factors).- 

• Soil depth and te.xture 

• Species composition 

• Competition 

• Physiologic age 

• Topography 

• Climate trends, past events 

• Air pollution 

These are longstanding conditions that predispose trees to effects that we will discuss. But one in 
particular, physiologic age. is different from chronological age. An 80-year-old tree is not an 80-year-old 
tree; it depends on where it is growing. An 80-year-old tree on a poor quality site, or a low productivity 
site, such a site index of 60, is more mature physiologicalK than that same age tree growing on a more 
productive site, say with a site index of 80. And we use this in modeling work to predict where oak 
decline is likely to be a problem. 

The second group of factors are the inciting factors (Sinclair 1965): 

• Defoliation 

• Drought 
Frost 

Stand disturbance 

• Air pollution 



24 



These factors are relatively short tenn, occurring at a point in time or a period of time that can be 
identified with the inciting event. And defoliation, spring defoliation in particular, is an important factor 
here. What happens with spring defoliation is that the carbohydrate chemistry of the tree is altered. Food 
is stored in roots as starch. In times of stress, such as when the crown is removed, the tree has to mobilize 
that starch into sugars. 

Finally, there are the contributing factor (Sinclair 1965). such as: 

• Root pathogens 

Armillaria root disease 

• Canker pathogens 

Hy poxy Ion 
Shoot cankers 

• Boring insects 

2-lined chestnut borer 
Red oak borer 

Root diseases, for example, can take advantage of a tree weakened by inciting factors through recognizing 
chemical changes in the roots and then switching from a saprophytic to a pathogenic relationship with the 
tree. These include Armillaria root disease, and in particular /lr/;7///t/r/V/ me/lea.. 

Using FIA data points of various dissections of forest type, the oak forest type is the most common one in 
the East, of course, and the message is that "there sure is a lot of oak out there" (Figure 4). When plots 
are displayed that are 'vulnerable", meaning that they have a relatively high basal area of oak, these are 
really saw timber and pole timber stands that have a high concentration of oak (using size as a surrogate 
for age). Vulnerable plots are concentrated in the Appalachian Mountains, the Blue Ridge in Virginia, the 
Eastern and Western Highland Rims in Tennessee, and the Ozark Mountains in Arkansas. "Affected" 
stands in Figure 4 are those in which oak decline symptoms are actually present, and these re\eal a 
pattern. There is about 3.6 million acres of oak forest type in the 12 southern states of this region, about 
10 percent of the total in the East. Oak and oak decline are especially abundant in the southern 
Appalachians, where chestnut would have been concentrated. 



Host Type 

, — ■— ^ 


y^^^ 


Vi 


ilnerable 




i'?^ 


Affected 




■ 




i 






\ 


I h 



Figure 4. FIA oak decline analysis for the USDA Forest Service's Southern Region, 
inventory cycle. 



1984-1989 



We mentioned earlier that defoliators have an important impact. The fall canker worm is a common 
defoliator, but probably made more serious by the loss of chestnut and its replacement with oak, as their 
favorite food is oak leaves. There was an outbreak of fall canker worm on some 1 0,000 acres on the Blue 
Ridge Parkway a couple of years ago that resulted in some tree mortality. This is an example of why 
spring defoliation is important in the oak decline scenario and in general tree health. Oaks produce an 
instant crown in the spring. If something comes along and removes those leaves in the first few weeks, 
then the tree has to make a decision about replacing that foliage, and starch needs to be converted to 
sugars from the roots. Spring defoliation stimulates a refoliation before the starch can be replenished, and 



25 



then you get the root diseases coming in. And the lesson also is that compounding stresses such as 
defoliation in combination with drought unhappily occur together frequently. Nitrogen content in leaves 
go up in drought periods, which makes it more palatable to insects, a positively-reinforcing loop. When 
predisposed oaks of advanced age are defoliated in the spring, combined with drought, disaster is waiting. 

Another added element is the gypsy moth, a non-native defoliator. The male has feather, antennae and 
the sex pheromone is from the female, which doesn't fly. Unhappily, the gypsy moth prefers oak species 
as host; they love to eat oak leaves. Among the more resistant species is the dearly departed American 
chestnut, and there is another array of hosts that are also relatively immune (Table 2.). Some other 
immune hosts are species that we do not need necessarily need more of. The bottom line is that the 
replacement of chestnut, a relatively resistant host to the gypsy moth, with the much more preferred oak _ 
again has forest health implications, especially in the oak decline scenario. 



Table 2. Tree host preferences for gypsy moth. 



Gypsy Moth Preferred Hosts 


Gypsy Moth Resistant Hosts 


Gypsy Moth Immune Hosts 


Oak species 


American chestnut and beech 


Ash 


Basswood 


Cottonwood and sourwood 


Fir 


Sweetgum 


Sweet and yellow birch 


Grape and holly 


Serviceberry 


Hemlock and pines 


Black locust 


Hornbeam, hop-hornbeam 


Blackgum and buckeye 


Sycamore 


Willow 


Walnuts and hickories 


Yellow-poplar 


Apple 


Black cherry and elms 


Striped maple 


Aspen 


Cucumbertree and sassafras 


Dogwood 


Gray, paper, and river birches 


Red and sugar maple 


Mountain-laurel 



Outbreak frequency, severity, and periodicity' tend to be different between native and non-native 
defoliators, and this has forest health implications. Outbreaks of non-natives tend to be more severe and 
have shorter return intervals than outbreaks of native species. 

So to summarize the chestnut blight-oak decline-gypsy moth interactions, we have an introduced 
pathogen superimposed on an altered forest due to the loss of chestnut and replacement with oak. That is 
an oversimplification; oaks weren't the only species to come in, but the\ were a ver\ significant 
component to replace chestnuts. But when you impose the interacting factors of oak decline, gypsy moth, 
forest composition, ana existing composition, oaks will decrease as a result of oak decline. This is 
somewhat site specific. Sometimes oaks replace themselves, but often they don't. The usual case is that 
there is incomplete oak replacement. So when a forest has 40-60 percent oak prior to these disturbances, 
you may end up with 20-25 percent remaining afterwards. Nobody projects that oaks will he lost 
completely; you couldn't get rid of them if you wanted to. There is an increase in the taller, mid-stoiy 
species, and it is the same scenario as what happened when oaks were positioned to take newly available 
space when the chestnut went out. You get shade tolerant mid-story species as a result of going decades 
without disturbance (no fire, no cutting, or very little anyway). You have a build-up in the mid-stor> of 
shade tolerant species like red maple, blackgum. and sourwood. Of course, this does not matter if all you 
want is something green out there. But if \ou place differential \alue on different species, then this could 
be a bad result, especially with regard to wildlife habitat components. 



26 



Evidence of forest composition change 

Unpublished data from the USDA Forest Service's Forest Inventory and Analysis (FIA) unit was 
assembled by Bob Anderson, recently retired. This was a study of a cluster of counties in northern 
Virginia where gypsy moth, oak decline, and dogwood anthracnose have come together over a number of 
decades. Between 1977-92, approximately three inventory cycles, there was a major change in trees 1 7 
inches and larger in diameter at breast height (Figure 5). The bottom line is that the large-tree component 
increased dramatically, especially for eastern white pine but other species, also. That is a positive change. 
But the picture is ver>' different at the other end of the size spectrum. In the trees 1-5 inches in diameter, 
which are going to be the next forest, over the same period of time, eastern hemlock showed a fairly 
robust increase but all other species declined. At the bottom, with the most negative changes, were the 
oak species. So the next forest is probably going to have a smaller oak component. 



Yellow Poplar ■ 

Scarlet Oak 

Beech 

Ash 

Virginia Pine 

Hemlock 

N Red Oak 

Red MapM 



^JESiSiiffijSaftMSE 



1 tT^Z-tA f i [ i I J 



Elm 

' Hickory 

VA Pine 

able Mtn Pine 

Bl Locust 

Pitch Pine 

Vhite Oak 

I Red Oak 

Black Oak 

;Chestnut Oak 

Scarlet Oak' 



IfiENtCHANGE 



Figure 5. Forest composition in northern VA, 
change in trees 17+" d.b.h., 1977-92. 



Figure 6. Forest composition in northern VA, 
change in trees 1.0-4.9" d.b.h., 1977-92. 



All of these changes have consequences for wildlife habitat. The mast quality and quantity is reduced, 
and this has consequences not only for food for mast-loving wildlife, but also in oak regeneration 
opportunities. This must be put in the context of oak as an incomplete surrogate for chestnut, and what 
chestnut provided in decades past. We have an increase in small openings, not such a bad thing in some 
contexts, depending on which wildlife species you are talking about and the landscape you are dealing 
with. There will be a change in species composition, both in the abundance and diversity of oaks, since 
the red oak group is more susceptible to the decline mortality than is the white oak group. Reduced 
canopy density, an increase in denning sites for types of wildlife, and structural changes from dead and 
downed wood, standing snags and so forth could be a good thing. But how many dead snags do you need 
in a landscape before they can no longer be exploited by the available wildlife populations? We tried to 
model what the effect of oak decline would be on acorn production. If all standing trees were alive, 
healthy, and producing an average amount of mast per year, the annual mast production would be 
somewhere on the order of 280 pounds per acre. But, of course, many of those trees aren't alive. A real 
stand was modeled in Virginia, on the Deerfieid Ranger District on what is still the GW Jefferson 
National Forest. Mast production from the dead oak was, of course, zero, and some trees had partial 
crown dieback and partial reduction in their mast-production capacity. Instead of 280 pounds per acre, 
the stand was producing 1 68 pounds on average. Projecting the current pace of decline, knowing that red 
oaks decline faster than white oaks, we predict that within 10 years of this inventory there will be only 
1 1 5 pounds per acre. Again, superimpose this on the context of a chestnut forest prior to its loss and 
replacement with oak. We don't have an accurate number for the mast production of chestnut historically 
on this kind of a site, but it might have been measured in tons per acre rather than hundreds of pounds per 
acre. 



27 



Sudden oak death and chestnut 



How does sudden oak death, or the potential of sudden oak death, t'lt into this? 1 tell people that there is a 
wide spectrum of possible outcomes with sudden oak death, from a chestnut blight t>pe of scenario to 
innocuous. Sudden oak death was confined to the West Coast (and Europe) until March of this year. It 
wasn't in the East until the disease (caused by Phytophlhora raniuruni) was shipped on nurser\ stock to , 
virtually all of the states plus Puerto Rico and the Virgin Islands. However, introduction does not 
neccssariK mean establishment. So what does sudden oak death look like? The diagnostic symptom is a 

bleeding stem canker on 
oak. but there are a lot of 
agents that cause cankers 
on oak stems. So 
bleeding cankers are not 
strictK speaking 
diagnostic, but a good 
clue. The bleeding is a 
running, wine or 
burgundy colored ooze 
(Figure 7). Underneath 
the bleeding spot are 
irregular lesions. On 
other species, P. 
nvuorum infection ma\ 
cause onl\ shoot dieback 
(madrone) or leaf blight 
(California-laurel). It has 
been said that sudden oak 

death is neither sudden, doesn't affect only oaks, and doesn't always result in death. So maybe that is not 
a good name. But it has crept into the common usage. You would have to say at the low end of the 
symptom scale that it might be 5 or 10 years from the infection to mortality. We haven't been looking 
long enough to know if some trees can recover. It doesn't appear so. 




Figure 7. Phylophihora ranioruni diseases - bleeding stem canker, shoot 
dieback, and leaf blight (clockwise from left). 



Prior to March 2004, the distribution of P. ramorum in North America was thought to be confined to the 
West Coast, to 12 central coastal California counties plus Curry Count\, Oregon, just north of the 
California border and a couple of hundred miles north of the most northerl> known site in California. We 
tried models to guide our survey efforts, to have a risk-based surve\ and to focus our resources in places 
where we were most likely to find this disease. We looked at climatic variables where the disease exists 
on the West Coast, and combined those with distributions of known potential hosts. As Figure 8 shows, 
there appears to be a heavy risk of sudden oak death in the southern Appalachians. 

On March 10, 2004, it became known or confirmed that P. ramorum pathogen was present in the 
Monrovia nursery in Los Angeles, California. Shipments of nurser\ stock from Monrovia, and another 
nurseiy called Specialty' Products, to eastern destinations ma\ have contained infected material. /'. 
ramorum has been confirmed in nursery stock sent to Mar> land, Virginia, North Carolina, Tennessee, 
Georgia, Florida, Louisiana, and Texas. Testing is continuing. Just because states are not known to have 
the shipments yet, does not mean it's not there. It is just that the testing is still underwas in many of those 
places. 



28 



Preliminary SOD Risk/Hazard Map 




/■'-. 



X 



(^ ■ i 



^ 






^- 



'^ \3f?¥^^&^ 






Relative Risk 

IHigh 
I « I Moderate 

rniow 



'^Z/ 



\ . \ ^'-> 

\ • \ 



^ 



,y 



-■^ 




^-^ 



V 



WOS/10Oct02 



Figure 8. Preliminary sudden oak death risk/liazard map. 



LITERATURE CITED 

Metcalf. H. 1912. The chestnut bark disease. P. 363-372 in Yearbook of the department of agriculture for 
1912, Washington, D.C. 

Metcalf, H., and J.F. Collins. 1909. The present status of the chestnut bark disease. USDA Bull. 141 part 
5, Washington, D.C, p. 45-53. 

Smith, D.M. 1976. Changes in eastern forests since 1600 and possible effects. P. 1-20 in Perspectives in 
Forest Entomology, Anderson, J.F., and H.K. Kaya (eds.). Academic Press, New York. 

Wicklum, D., and R.W. Davies. 1995. Ecosystem health and integrity? Can. J. Bot. 73:997-1000. 



29 



30 



Steiner, K. C. and Carlson, J. E, eds. 2006. Restoration of American Chestnut To Forest Lands - Proceedings of a 
_^ Conference and Workshop. May 4-6, 2004, The North Carolina Arboretum. Natural Resources Report 
NPS/TJCR/CUE/NRR - 2006/00 1 , National Park Service. Washington, DC. 



CURRENT STATUS OF CHESTNUT IN EASTERN US FORESTS 

Wiliiam H. McWilliams', Tonya W. Lister', Elizabeth B. LaPoint', Anita K. Rose", and John S. Vissage^ 
USDA Forest Service, Northeastern Research Station, Newtown Square, PA 19081 USA 

(wmcwilliams@fs.fed.us) 

- USDA Forest Service, Southern Research Station, Knoxville, TN USA 37919 

^ USDA Forest Service, North Central Research Station, St. Paul, MN USA 55108 



Abstract: The USDA Forest Service, Forest Inventory and Analysis (FIA) program provides the 
opportunity to assess the current distribution of American chestnut {Castanea dentata (Marsh.) Borkh) 
and prospective trends. Assessing chestnut using the FIA data was challenging because of the coarse 
nature of the FIA sample and chestnut's rarity in natural forests: however, a basic analysis of location and 
character provide important infonnation for scientists seeking to re-establish chestnut. Chestnut occurred 
from Vermont to Alabama or from roughly 45° to 30° north latitude. The estimate of the area of forest 
land with chestnut at least 1.0-inch in diameter was 2.8 million acres. The area with the highest 
concentration of chestnut aligned well with Braun's oak-chestnut forest region. About two-thirds of the 
chestnut sample was on private land and 87 percent was found in oak-hickory stands that vary 
considerably in composition from north to south. Derivation of a population estimate for the total number 
of chestnut stems was precluded by missing data. Trends in the existing sample of sapling and tree-size 
stems suggest a decrease in sapling-size stems and an increase in tree-size stems. Future research on 
chestnut using FIA data could include filling in data gaps as new inventories are completed, development 
of improved indicators using new national core health variables, and analysis using geographic 
information systems (GIS). 

Keywords: American chestnut / Castanea dentata I distribution / map / Forest Inventory and Analysis / 
oak-chestnut forest region. 



INTRODUCTION 

American chestnut {Castanea dentata (Marsh.) Borkh.) is still a component of the forest understory in 
much of its native range despite its extirpation as an overstory component by the chestnut blight 
{Endothia parasitica (Murr.) Anders and Anders) (Ciyphonectria parasitica (Murrill) Barr.) beginning in 
the early 1900"s (Paillet 1988, Stephenson and Adams 1991 ). The USDA Forest Service, Forest 
Inventory and Analysis (FIA) program conducts large-scale forest inventories across the United States 
and provides the opportunity to assess the current distribution of chestnut and prospective trends. 
Assessing chestnut using the FIA data was challenging because of the coarse nature of the FIA sample 
and chestnut's rarity in natural forests. However some basic analyses of location and character can 
provide important information for scientists seeking to re-establish chestnut. An examination and 
analysis of available data is provided, along with cautionary comments on data interpretation. 



MATERIALS AND METHODS 

In 1999, the FIA program converted from a periodic system, in which states were inventoried every 10 to 
15 years, to an annual system with fixed portions of a state's forests measured annually. FIA uses a three- 
phase system to inventory and monitor forests. Phase 1 uses remote sensing to stratify the land base as 
forest and nonforest and assign a representative number of acres to each sample plot measured in Phase 2. 



31 



Phase 2 consists of field measurements collected on a grid of sample plots spread across the United 
States. Each plot is made up of four 24-foot circular fixed-radius subplots for inventor) of trees at least \ 
5.0 inches in diameter. Trees less than 5.0 inches are inventoried on 6.8-foot circular fixed-radius micro- 
plots nested within each subplot. At each sample plot, a suite of plot and tree-le\el measurements are 
collected. Each Phase 2 plot represents about 6,000 acres, although some states have intensified sample 
grids. Phase 3 measurements are collected on a limited number of Phase 2 locations and include more 
detailed forest-health parameters, such as tree crown condition. 

The Phase 2 sample data were used to identify locations where chestnut occurs and to characterize sites, 
stands, and tree sizes. Tree-size class provides a surrogate for age or stage of development. Seedlings are 
trees that are less than 1 .0-inch in diameter and at least 0.5 and 1.0 feet in height for coniferous and 
deciduous species, respectively. Saplings range from 1 .0 to 4.9 inches. Tree size is defined as 5.0 inches 
in diameter and larger. The population estimate of the total forest land acreage with chestnut is 
mentioned; however, it should be recognized that chestnut's occurrence is rare and discontinuous, so the 
accuracy and precision of population estimates and related findings are often low. 

Other sampling issues associated with chestnut may affect estimates and conclusions. Misidentification 
can occur because of confusion with Asian chestnuts, cultivars, and similar species. Alleghen\ chinkapin 
{Caslcmea pinnila (Mill.)) shares much of the current distribution and may have resulted in errors of 
inclusion. Also, when tallying clumps of seedlings of a single species. FIA crews usualK record the most 
dominant stem. In some older inventories, chestnut may have been grouped into a nonspecific species 
code. In other cases, seedling tallies were limited to the four most dominant species and only collected if 
no larger trees occurred on the sample plot. The lack of seedling data for all states was the major 
limitation of the study. Other less significant factors include differing sample grids, plot designs, and 
methods of measuring snags among states and inventory dates. 

Data were screened for obvious outliers. Less obvious or questionable plots were allowed to remain in 
the dataset, recognizing that the distribution maps based on these data may contain errors of inclusion or 
exclusion. Errors of exclusion are often known. For example, FIA field staff in Pennsylvania reported 
many sightings of chestnut in the vicinity of sample plots, but chestnut was not actually sampled. 

Despite these difficulties, FIA data are the only source of consistently gathered sample data on the 
contemporary occurrence of chestnut throughout its original distribution. Maps of chestnut distribution 
and related stand characteristics provide useful information for scientists interested in location and extent. 
More specific local results are available through herbarium studies and other monitoring. 

Sources of FIA data used to characterize chestnut came from all available digital data for the states within 
the natural range of chestnut prior to the blight (Little 1977). This included data from the older periodic 
inventories and the new annual inventories for states w here chestnut appeared in the inventory (Table 1 ). 
The most current inventories occurred from 1991 to 2002 and previous inventories from 1980 to 1995. In 
some cases, only one inventory was available. In order to minimize the amount of error introduced, 
annual inventoiy data were used only if at least 50 percent of the sample plots in an\ given state had been 
measured. Although this allowed the most current data to be used, some imprecision was apparent in the 
results. Other source data are contained in the numerous state-level reports published b\ FIA since the 
1930s, but documenting the significant post-blight decline of the early and mid-1900s went beyond the 
objectives of this study. As such, the analysis covered the current resource and the latest trend 
information available. 



32 



Table 1. Sources of FIA data used to characterize contemporary occurrence and distribution of American 
chestnut in the eastern United States. 







Previous 


Inventory -- 








— Current 1 


Inventory — 




-- ^— -^ 






Number 


Number 








Number 


Number 


.N .^ 






of 


of plots 








of 


of plots 






Inventory 


forested 


with live 






Inventory 


forested 


with live 


State 


Year 


type 


plots 


chestnut' 


Year 


type 


plots 


chestnut' 


Alabama 


1990 


Periodic 


3923 


3 


2000 




Periodic 


4421 


3 


Connecticut 


1985 


Periodic 


215 


2 


1998 




Periodic 


319 


2 


Georgia 


1989 


Periodic 


7713 


2 


1997 




Periodic 


7272 


4 


Illinois 


1985 


Periodic 


1169 





1998 




Periodic 


1750 


2 


Indiana 


1998 


Periodic 


1605 


1 


1999- 


2002 


Annual 


738 





Kentucky 


1988 


Periodic 


2005 


4 












Maine 


1995 


Periodic 


2733 


1 


1999-, 


2002 


Annual 


2560 





Maryland 


1986 


Periodic 


716 


3 


1999 




Annual 


562 


8 


Massachusetts 


1985 


Periodic 


243 


1 


1998 




Periodic 


583 


14 


Michigan 


1993 


Periodic 


10849 





2000-2002 


Annual 


4200 


1 


New 




















Hampshire 


1983 


Periodic 


590 


4 


1997 




Periodic 


853 


4 


New Jersey 


1987 


Periodic 


254 


2 


1999 




Periodic 


429 


4 


New York 


1993 


Periodic 


3063 


14 












North Carolina 


1984 


Periodic 


5676 


37 


1990 




Periodic 


5965 


31 


Ohio 


1993 


Periodic 


1802 


1 












Pennsylvania 


1989 


Periodic 


3208 


53 


2000- 


2002 


Annual 


1929 


19 


Rhode Island 


1985 


Periodic 


61 





1998 




Periodic 


123 


6 


South Carolina 


1993 


Periodic 


4563 


1 


1999-2001 


Annual 


2815 





Tennessee 


1989 


Periodic 


2315 


4 


1999 




Periodic 


2838 


9 


Virginia 


1992 


Periodic 


4424 


45 


1998- 


2001 


Annual 


3169 


41 


West Virginia 


1989 


Periodic 


2628 


11 


2000 




Periodic 


2188 


21 



At least I.O-inches in diameter. 



Current Distribution 



RESULTS 



Chestnut samples plots were found between about 45° and 30° north latitude, but 85 percent were between 
41° and 35° north latitude. Figure 1 depicts sample plots where live or dead chestnut trees at least 1 .0 
inch diameter were present in any of the inventories since 1980. Plots containing only dead trees were 
included to provide the most inclusive range description possible. Chestnut occurred from Vermont to 
Alabama and from Illinois in the west to Maine in the east. It was native to Ontario also, but FIA data do 
not cover Canada. 

The estimate of forest land area with live chestnut at least 1 .0 inch in diameter is 2.8 million acres. This 
estimate is based on the most recent cycle of inventories. The FIA definition of forest land includes areas 
at least 1 acre in size, at least 10 percent stocked with trees (or has been in the past), in a strip at least 120- 
feet wide, and not characterized by land uses that inhibit nonnal forest regeneration and succession (such 
as mowing). As such, some land with chestnut trees in fencerows or other land with trees would be 
excluded. It should also be noted that the estimate of forest land with chestnut would be higher if 
seedlings were included in the analysis. 



33 




•— 




tu 








3 




c 




:- 




r" 








■^^ 




T3 


^ 




^^ 


^ 


U"i 


^ 


o 


<B 


— 


o 


•z. 


u. 


~ 


O 


c: 


-* 


t. 


^ 


r<^ 


^. 


■*■ 


<U 




a> 


^ 


L 






^ 


■*-' 


D/J 


.^ 


O 


^ 


f^ 






/: 


, . 


O 


■J-. 


r^ 


(L» 


CJ 


;_ 




^> 














^W 




o 


t^ 






^ 




CJ 


^ 


r- 


y. 


^ 


O 


< 




•o 


U 


(S 


ji 


a> 




T3 


,^v 


;_ 


^^ 


C 


o 






O 


""^ 


> 


■j: 












5 

■r. 

1) 


■K- 


^ 


^ 




CA 


^ 


■*— ' 




r- 


r: 


^— 


o 


c 


:— 










_o 




o 










^^ 


*"^ 


C3 






•y) 






;/: 


< 








LjL 


c 


"U- 








■— ' 


^_^ 


^ 






1) 


_ ^ 


CJ 


•^ 


^ 






;C 


X 


^ 


-o 




o 


cr 






CJ 


C 






-o 




n 




o 




o 




c« 




.2 


^ 


"C 




Q 


O 




u. 


f~ 


3 


5 


iij 


> 






u. 


• — 



34 



The top six states by number of sample plots with live trees at least 1 .0-inch are Pennsylvania (53), 
Virginia (45). North Carolina (37), West Virginia (21), Massachusetts (14), and New York (14). Note 
that this ranking is based on a sample slightly different than the one depicted in Figure 1; this ranking is 
based on the most recent full periodic inventory of each state to utilize the largest sample possible. As 
such, larger states have larger samples. 

High concentrations of chestnut were found in southern New England (Fig. 2), Pennsylvania, West 
Virginia-Virginia-Maiyland (Fig. 3). and east Tennessee and western North Carolina (Fig. 4). Figures 2 
to 4 also show the distribution of samples by tree-size class. It is notable but not surprising that the 
current extent and abundance suggested by this somewhat fragmented picture aligns with Braun's (1950) 
oak-chestnut forest region and mixed Mesophytic region to the west. 

Site and Stand Characteristics 

FIA site and stand data were used to characterize forest stands containing chestnut. The basic findings 
were similar to those of Braun (1950) who described the extent and abundance following the blight using 
existing reference sites and standing dead trees. 

The contemporary chestnut population occurs across a range of slopes, but was rare on the steepest slopes 
(Table 2) and most common at elevations below 2000 feet. Chestnut was most prevalent on northeast- 
facing slopes, but was common on all aspects. About three-fourths of the chestnut occurred on mesic 
sites. It was also found on xeric sites. Chestnut was rare on very wet sites. 

Forest resources in the eastern United States are primarily controlled by private forest landowners and 
stands containing chestnut are not an exception. Nearly 60 percent of the chestnut sample was on private 
land. The private owner group is a complex mix. from timber industry to small family owners. The 
National Forest System was the second most common owner with 25 percent of the total chestnut sample. 
The other public sample is incomplete because the existing sample excludes National Park Service land in 
the North Carolina portion of the Great Smoky Mountain Park and the Adirondack and Catskills State 
Parks in New York. New inventories will cover these lands in the future. 

As was similarly found by Braun (1950), chestnut occurred most commonly in the oak-hickory forest- 
type group. The oak-hickor>' group covers a wide range of forest cover types, primarily of mixed-oak 
composition with varying proportions of other associates, such as yellow-poplar {Liriodendron tulipifera 
L.), hickory {Carya sp.), and other species depending on the region. More than 80 percent of the sample 
plots with chestnut were found in oak-dominated stands. Although found across the East, high 
concentrations of oak-dominated stands are \Qvy common in the Appalachian Mountains from central 
Pennsylvania to northern Alabama (McWilliams and others 2002). 

In southern New England, chestnut occurs on glaciated soils with higher occurrence of red maple {Acer 
rubnnn L.), sugar maple {A. saccharum Marsh.), beech {Fagus gnindifoUa Ehrh.), and white ash 
{Fraximis americana L.) than regions to the south. The northern Ridge and Valley region of 
Pennsylvania is characterized by mixed-oak species with yellow-poplar and hickor>' being relatively rare. 
Yellow-poplar and hickory become more common in stands containing chestnut in the southern tier of 
Pennsylvania and the northern Blue Ridge areas of West Virginia and Virginia. Further south in the 
Southern Appalachians and Great Smoky Mountains, the number of associates increases. These 
differences in associates emphasize the high degree of species heterogeneity that exists throughout eastern 
North America. 



35 



r^y 




^ 



Plot Density Index 
^1 Free | | 1 - 5 

I I Sapling [^ 6-10 



Seedling 



11 + 



Figure 2. Distribution of live American chestnut showing chestnut plot densit> index and proportion of 
seedlings, saplings, and trees by county in Southern New England. [Note: Plot densit\ index = 
number of plots with live chestnut/county area (sq. mi.) * 1000]. 




(b 



Plot Density Index 
^1 Tree [^ 1 - 5 

I I Sapling | | 6-10 
I I Seedling ^| 11 + 



Figure 3. Distribution of live American chestnut showing chestnut plot density index and proportion of 
seedlings, saplings, and trees by count\ in Pennsylvania. Mar\ land. West Virginia, and 
Virginia. No seedling data available for Virginia. [Note: Plot density index = number of plots 
with live chestnut/county area (sq. mi.) * 1000]. 



36 




V ^m Plot Density Index 
H Tree [^ 1 - 5 

I I Sapling |^ 6-10 
Hi 11 -•■ 



Figure 4. Distribution of live American chestnut showing chestnut plot density index and proportion of 
saplings and trees by county in east Tennessee and western North Carolina. 
[Note: Plot density index = number of plots with live chestnut/county area (sq. mi.) * 1000]. 



The distribution of forest land containing chestnut by stand-size and stocking class was similar to the 
distribution for all forest land across its range. By stand-size class, the distribution was 60 percent 
sawtimber size, 28 percent mid-size, and 12 percent sapling-seedling size. Sawtimber stands are 
dominated by trees at least 9.0 and 1 1 .0 inches in diameter for coniferous and deciduous species, 
respectively. Mid-size stands are dominated by trees at least 5.0-inches in diameter but smaller than 
sawtimber size. Sapling-seedling stands contain mostly trees less than 5.0 inches. Eighty-eight percent 
of the forest land with chestnut was in either medium (30-69 percent stocked) or fully (70-100 percent) 
stocked stands. 

Large areas of eastern U.S. mountain and upland forests are evolving along similar compositional and 
structural trajectories. Stands containing chestnut are representative of these conditions. Dominant trends 
include forest land with increasing numbers of large-diameter trees, decreases in small to mid-range trees, 
mismatches between overstory and understory species composition, relatively few young sapling-seedling 
stands, and often, regeneration difficulties. Susceptibility to prominent pests, such as Asian long-homed 
beetle, elm-ash borer, and hemlock wooly adelgid threaten many of the canopy dominants that occur over 
significant areas. Future developments in these forests will affect chestnut's niche within natural and 
disturbed forest land. 



37 



Table 2. 



Site and stand characteristics expressed as a percent of plots with live American chestnut 
(at least I.O-inches in diameter) sampled in the most recent FIA inventories conducted in 
the eastern United States. 







— Percent 


of Sample Plots — 






Slope 




Elevation 




Aspect 






0-5% 12 




500' 


8 


NE 




33 


6-10% 10 




1000" 


38 


SE 




20 


11-20% 19 




1500' 


17 


sw 




22 


21-30% 14 




2000' 


14 


NW 




24 


31-40% 15 




2500' 


11 








41-50% 14 




3000" 


6 


Moisture Class 




51-60% 10 




3500' 


6 








61-70% 4 




4000' 


1 


Hydric 




- trace - 


71-80% 2 




- 




Mesic 
Xeric 




73 
27 


Ownership 








Stand Size 






National Forest 


25 






Sapling- Seedling 


12 




Other Federal' 


2 






Mid-size 


28 




Other Public 


14 






Sawtimber 


60 




Private 


59 












Forest Type Grouping 






Relative Stocking 






White Pine-Hemlock 


2 




Over (> 100%) 


8 




Spruce-Fir 




-trace- 




Full (70-100%) 


55 




Loblolly-Shortleaf 




2 




Medium (30-69%) 


33 




Oak- Pine 




5 




Low (< 30%) 


3 




Oak-Hickor> 




84 










Elm-Ash-Cottonwood 


1 










Northern Hardwoods 


6 










Aspen-Birch 




1 











' The Other Federal ownership excludes forest land with chestnut in the Great Smoky Mountain National Park and 
the Adirondack and Catskills State Parks. 



Numbers of Stems 

Derivation of a population estimate for the total number of chestnut stems was precluded by some 
missing information for seedlings. However, an examination of the existing sample of sapling and tree- 
size stems suggests a decrease in sapling-size stems and an increase in tree-size stems over the two most 
recent inventory cycles. It is unfortunate that the seedling sample was incomplete because structural 
trends are not completely discemable. The critical question is what degree the seedling/sprout resource is 
changing. This resource represents recruitment of future chestnut stems. While the increase in tree-size 
stems is encouraging in terms of viability, the long-term sustainability of chestnut depends on rccmitment 
of chestnut stems. 



38 



CONCLUSIONS 

Tlie FIA data provide a coarse description of the chestnut resource as it occurs in today's forests. The 
data indicate that the existing population of chestnut occupies the core of the oak-chestnut forest region 
described by Braun (1950) with rehc communities found across its original range described by Little 
(1977). This is not surprising because chestnuts exist today mainly as sprouts (Paillet 1988). It is not 
possible to make a conclusive statement of the long-term sustainability of chestnut due to limitations of 
the current dataset. Future inventories will fill existing gaps and provide additional data needed for more 
thorough analysis of structural changes and trends in spatial extent. A significant benefit of the new 
national FIA system is improvement to the seedling and sapling measurement protocols. All seedlings 
are now tallied in a consistent manner. Sapling measurements now include total height, crown class, and 
condition. Remeasurement of these parameters will lead to improved datasets over the next 5 to 10 years. 

Future extensions of research on chestnut using FIA data are readily apparent. The most obvious need is 
to provide more comprehensive data for analysis. Pending release of FIA results for Kentucky, North 
Carolina, and New York will fill some critical needs. The inclusion of seedling information in FlA's 
current national protocols will be particularly helpful. Once the gaps in data are filled, a more complete 
analysis of site occupancy could be conducted using geographic infomiation systems (GIS). Modern GIS 
software is capable of analyzing hundreds of data layers that could help delineate characteristics 
associated with chestnut's occurrence. Improvements to FIA Phase 2 and new Phase 3 variables have 
resulted from nationalization of FIA protocols. For example. Phase 3 includes tree crowns, damage, 
down woody material, and others. These new variables offer the opportunity to develop improved 
indicators of chestnut condition and extent. The opportunities for improvement of our knowledge of 
chestnut at the landscape level are immense. 



LITERATURE CITED 
Braun, E.L. 1950. Deciduous forests of eastern North America. Hafner Publishing, New York, NY. 596 p. 

Little. E.L., Jr. 1977. Atlas of United States trees, vol. 4. Minor eastern hardwoods. USDA Misc. Publ. 
1342. Washington, DC. 17 p. 230 maps. 

McWilliams, W.H., R.A. O'Brien. G.C. Reese, and K.C. Waddell. 2002. Distribution and abundance of 
oaks in North America. P. 13-33 in Oak forest ecosystems: ecology and management for wildlife, 
McShea. W.J. and W.M. Healy (eds.). John's Hopkins University Press, Baltimore and London. 

Paillet, Frederick L. 1988. Character and distribution of American chestnut sprouts in southern New 
England woodlands. Bull. Torrey Bot. Club 1 15:32-44. 

Stephenson, S.L., H.S. Adams, and M.L. Lipford. 1991. The present distribution of chestnut in the upland 
forest communities of Virginia. Bull. Torrey Bot. Club 1 18:24-32. 



39 



40 



Steiner, K. C. and Carlson, J. E. eds. 2006. Restoration of American Chestnut To Forest Lands - Proceedings of a 
Conference and Workshop. May 4-6, 2004, The North Carolina Arboretum. Natural Resources Report 
NPS/NCR/CUE/NRR - 2006/001, National Park Service. Washington, DC. 

CHESTNUT AND WILDLIFE 

Frederick L. Paillet 
Department of Earth Sciences, University of Maine. Orono, ME 04469-5790 USA (fpaillet(??maine.edu) 



Abstract: The interaction of chestnut with wildlife can be separated into two issues: 1) Chestnuts as a 
resource for wildlife consumption; and 2) Wildlife as a dispersal agent for chestnut seed. The chestnut 
mast resource was probably not qualitatively different from that of other large- fruited mast species. 
Chestnut mast may have been quantitatively better because chestnut crops were probably more reliable 
from season to season, and because chestnut has been partly replaced by tree species such as tulip poplar 
that do not produce nuts. These factors may result in a 20 to 50% increase in total mast production if 
chestnut is re-established as a canopy-dominant tree in American forests. Dispersal of chestnut seed and 
seed burial by wildlife were important because chestnuts have little resistance to frost or desiccation. 
However, chestnuts as seed exhibit few of the mechanisms usually associated with a trade-off between 
dispersal and predation. Various field studies and old forestry data confirm that chestnut reproduction 
occurred most often by coppice sprouting, and that chestnut seedlings were rare or absent in many 
chestnut stands. I propose that there is no selective pressure on chestnut for the deterrence of predators 
because chestnut reproduction is based on a combination of wildlife dispersal of nuts and long term 
survival of the few seeds that do manage to gemiinate. My studies show that many chestnut seedlings 
have survived for at least a century and possess root collar sprouting characteristics designed to insure 
that "old seedlings" remain juvenile essentially forever. Thus, chestnut reproduction by seed is heavily 
biased in favor of mechanisms to promote long-distance transport. The near immortality of established 
chestnut root stocks probably offsets any selective value in deterring seed predation within chestnut- 
dominated stands. 



INTRODUCTION 

American chestnut, Castanea dentata, was once one of the leading tree species in the forests of the 
Appalachians, Allegheny Plateau, and southern New England (Smith, 2000; figure 1 ). Mature chestnut 
trees were removed from these forest ecosystems after the introduction of chestnut blight in the New York 
City area sometime around 1900 (Anderson, 1974). Chestnut mast was completely removed from 
eastern forests, although numerous chestnut sprouts continue to exist and even flourish in the forest 
understory (Paillet, 2003). A related species, Allegheny chinquapin {Castanea pumilo), was also affected 
by blight, but chinquapin grows as a shrub or small tree, so that some fruiting probably continued to occur 
even after the appearance of blight within the range of that species (Paillet, 1993). I therefore conclude 
that blight effectively eliminated seed reproduction by chestnut, and drastically limited seed reproduction 
on the part of chinquapin. 

The elimination of a significant chestnut mast crop from American forests had two important effects: loss 
of an important food source for wildlife, and loss of the mechanism for seed dispersal on the part of the 
two native Castanea species. This report addresses the relationship between wildlife and chestnut by 
considering these distinct issues. The analysis begins by assessing what is known about chestnut ecology 
and the character of the chestnut mast crop in the years before blight arrived in America. These facts are 
the used to project how a restoration of that resource might affect wildlife in future forests. The 
projection is made possible by the general similarity of chestnuts to large acorns in terms of size, seed 
packaging, and nutrition (grams of carbohydrate, fat, and protein per kilogram, Krockmal and Krochmal, 
1982). The analysis then considers how wildlife predation on chestnut seed might alfect chestnut 



41 



dissemination and reproduction. This, in turn, requires consideration of both the physiology of chestnut 
production and the ecological factors related to chestnut seedling establishment in the wild. 




NATURAL RANGE OF 
AMERICAN CHESTNUT 

NATURAL RANGE OF 
CHlNQUAPtN 



KiLOMnens 



Figure 1 . The natural range of chestnut and chinquapin in North America (from Paillet. 2003) 



THE FACTS WE HAVE TO WORK WITH 

Although there are no natural American chestnut forests to study using the techniques of modem ecology, 
several different avenues of investigation provide some information about the role of chestnut in pre- 
blight American forests. These methods include: I ) Analysis of early forestr\ literature; 2) Pollen 
analysis from bogs and ponds that were once surrounded by chestnut forests; 3) Study of naturalized 
stands of American chestnut established beyond the range of blight and natural forests of related chestnut 
species in Russia and China; and 4) Reconstruction of conditions in former chestnut forests by 
examination of stumps and surviving sprouts. The known facts can be summarized as follows: 

Chestnut was a leading species in eastern North America, often approaching 50% of total stand basal area 
in upland forests from North Carolina to Connecticut (Nichols, 1913; Frothingham, 1912; Zon. 1904; 
Buttrick and Holmes, 1913). Chestnut was successfully propagating in natural forests both before and 
after the arrival of Europeans. 

Pollen studies demonstrate that chestnut was one of the last deciduous tree species to arri\c within its 
natural range in the Holocene from a glacial refuge probabK located on the southern coastal plain (Da\is. 
1969; Whitehead, 1979). The best estimate of the glacial habitat of chestnut includes the well-drained 
bluffs along rivers draining western Florida and Georgia (Watts, 1979). 



42 



Chestnut was only found growing on non calcareous soils, and was rare on soils developed on clay-rich 
glacial till (Russell, 1987). The survival of isolated chestnut trees on till soils in the Midwest 
demonstrates that the aversion to heavy soils is probably not the result of toxicity, but is caused by a 
combination of inability to compete and microsite conditions (Paillet and Rutter, 1989) 

Pollen studies show an abrupt expansion of chestnut at about 3000 years ago in New England (Brugham, 
1978; Whitehead, 1979; Foster and Zybryk, 1991 ). Pollen ratios shift from less than a few percent to 
20% or more. Chestnut pollen is under-represented in pollen profiles by a factor of 3 (Paillet et al, 1991) 
and yet relatively small chestnut pollen particles can be transported long distances. Thus it is unclear 
whether chestnut was present in low numbers in New England and upstate New York before 3000 years 
ago, or the small amount of chestnut pollen present before then resulted from long-distance transport. 
The 3000 year old shift in chestnut pollen corresponds almost exactly with an increase in spruce pollen in 
the same catchments, suggesting a climate change may be involved (Davis et al. 1980). 

The introduction of European land use practices apparently affected chestnut because many pollen 
profiles show a near doubling of the proportion of chestnut coinciding with the arrival of European 
settlers (Brugham, 1980). This is almost certainly a result of the conversion of bottomland soils to fields 
or pasture, and the changes in disturbance regime on remaining upland woodlots. 

Chestnut was destroyed as a forest tree by blight over the period 1900-1950 throughout its natural range 
(Anderson, 1974). All evidence suggests that the chestnut seed crop was completely removed from the 
forest as a source of food for wildlife and as a source of seed propagation for the forest (Paillet, 2003). 

Even though chestnut trees were destroyed, chestnut sprouts are abundant in modern forests. In fact, 
chestnut sprouts are so pervasive that they consistently show up a significant contribution to shrub cover 
(Adams and Stephenson, 1 983; Boring et al, 1981). 

Chestnut sprouts grow as fast as or faster than the sprouts from other competing species when released by 
disturbance (Adams and Stephenson, 1983; Stephens and Waggoner, 1980). Chestnut sprouts are 
recognized as the leading component of biomass in the years immediately after clear cutting in some 
Appalachian forests (Boring et al. 1981 ). This ability represents a strong adaptive advantage in "sprout 
hardwood" forests where canopy regeneration is dominated by coppice sprouts (Hibbs, 1983) 

Early historic forestry practices were based on coppice sprout regeneration (Buttrick and Holmes, 1913; 
Matoon, 1909; Smith, 2000). Older references clearly and repeatedly indicate that chestnut propagation 
from seed was not effective and that woodlots should be managed so as to encourage regeneration of 
chestnuts by stump sprouting. 

Chestnut has a significant range overlap with chinquapin. Both species reproduce by sprouting as well as 
by seed. The one major difference is that chestnut (a large forest tree) sprouts only from pre-formed buds 
on the root collar, whereas chinquapin (a large shrub or subcanopy tree) sprouts from an extend region of 
the lower stem and upper root system (Paillet. 1993). 



THE CURIOUS CASE OF THE ANCIENT SEEDLINGS 

Although much of the early forest literature and folklore refers to "chestnut trees smoldering at the roots", 
careful study of surviving chestnut spouts shows that almost all of these sprouts are old seedlings. The 
old forestry literature indicates that chestnut sprouts only from the root collar and not from roots at a 
distance from the stump as in the case of aspen and beech (Matoon, 1909; Zon, 1904). Chestnut wood is 
resistant to decay so that the remains of blight-killed trees can be recognized in the field (Saucier. 1973). 



43 



Thus, it is possible to determine whether surviving chestnut sprouts originated from former trees. Such 
analysis shows that almost all living sprouts never were attached to a canopy-dominant tree (Pailiet, 
1984). Most sprouts survive for an extended period as small upright trees 2-4 m in height, with an 
enlarged root collar covered with suppressed buds (figure 2). Maps of dense sprout population show that 
only a small number originated from the base of former trees (figure 3). 

These results indicate that the many living chestnut sprouts in modem forests are old seedlings that have 
survived for many years without ever becoming a large tree. The remains of large chestnut trees on the 
site illustrated in figure 3 were killed in 1922 according to the cross-correlation of ring widths with 
standard chronologies (Pailiet, 1984). In addition to simply surviving, the many chestnut sprouts retained 
their small tree form through several cycles of stem destruction by blight or mechanical damage. Pailiet 
(1993) suggests that this is not coincidence, and indicates this is an adaptation to insure that established 
chestnut seedlings remain "perpetually juvenile" (figure 4). This cycle of stem regeneration and root 
system abandonment appears designed to maintain seedling form indefinitely as a method of advanced 
regeneration. As a result of this sprouting mechanism, an established chestnut seedling might be able to 
assume its place in the canopy a century or more after germinating from a chestnut deposited in the forest 
litter. Although other American trees species are capable of producing coppice sprouts, the controlled 
release of stems and systematic replacement of the roots system in the chestnut seedling sprout cycle is 
unique to Castanea dentata. 



CHESTNUT AS A RESOURCE FOR WILDLIFE 

Chestnuts must have been a nearly ideal food for mast-consuming wildlife. Ripe chestnuts are not 
protected by a predator-resistant husk or shell, or by chemicals such as tannins. The general food 
"package" presented by chestnut was probably similar to that of the northern red oak or swamp chestnut 
oak (oak species producing relatively large fruit) in terms of fruit size and quantity of seed produced by 
an individual tree. The nutritional value of chestnuts (grams of carbohydrates, proteins, and fat per 
kilogram of nuts; Krochmal and Krochmal, 1982) is comparable to that of various oak species. Chestnut 
canopy crowns in naturalized stands of American chestnut are qualitatively similar to large oaks, and 
probably produce about the same number of nuts per branch tip as a northern red oak when the tree is 
producing at full capacity. These arguments suggest that a chestnut-dominated woodlot produced a nut 
crop that was qualitatively similar to the crop produced by an analogous stand of northern red oak trees. 

Restoration of chestnut to American forests might still have an effect on wildlife b\ making a quantitative 
difference in total mast crop. An increase in total mast could be produced b> two mechanisms: 1 ) 
chestnut seed crops would be more regular than those of other nut producing trees; and 2) chestnut might 
displace some tree species that do not produce nuts. Chestnut seed crops would exhibit relativeK few 
years of low production because chestnut flowers in early summer when there is no possibility of frost 
damage to ovaries and catkins, and because the nuts mature quickly so as to minimize exposure to insects 
or other damage to the immature fruit. This would produce a net increase in mast crop when averaged 
over several \ears, and could have a beneficial effect for wildlife in those \ears when a late frost 
diminishes other nuts. Chestnut is adapted for relative well-drained sites, and would displace mosth 
other nut producing trees like oak and hickorv on the drier end of the spectrum. However, chestnut can 
also grow on more mesic sites where it mixes with beech, tulip poplar, maple, basswood, and hemlock. 
Of these trees, only beech produces a significant mast crop. Together, these mechanisms suggest that 
reintroduction of chestnut might increase total nut production for wildlife by something like 20 to 50%. 



44 







Figure 2. Typical chestnut sprout living in the understory of former oak-chestnut woodland in New 
England; the log leaning against the stone wall represents the typical appearance in 1983 of a chestnut 
tree killed by blight in 1922. 



45 



w^ 




METERS 



EXPLANATION 
• CheiUiut sprout systems wi(h no obvious signs d attached 
root system from pre-bllghl tres 



/ 



Pre-blighl chestnut wood 
O Pre-blight chestnut stump (Cut 1 925-35) 
4: Chestnut sprout clearly attached to pre-blight root system 
^ Completety dead chestnut sprout clones 



Figure 3. Remains of ciiestnut trees killed by bligiit in 1922 and living chestnut sprouts present in 1983 
on a one-hectare plot in Andover. Massachusetts (from Paillet, 1984) 



46 



YEAR = 15 -YEAR = 20 




Figure 4. Schematic illustration of the chestnut seedling growth cycle where seedlings develop root collar 
buds that replace injured or senescent stems, generate a new root system and root collar buds, and retain 
juvenile growth form and defect-free stem base throughout an indefinite number of such stem 
replacement cycles. 



WILDLIFE AS A DISPERSAL AGENT FOR CHESTNUT 

Ecological literature assumes that mast crops are "designed" to be attractive to wildlife to provide 
effective dispersal of seed. Seed predators and trees interact in complex ways to insure that predators are 
"rewarded" for their role in planting nuts, while seed protection mechanisms insure that at least some nuts 
survive to germinate and grow. The seasonality of nut crops naturally results in seed caching by some 
predators to make the mast crop last the entire year. But this also allows other predators to develop 
rooting behavior to "harvest" these caches. Thus it is beneficial for nut-producing trees to develop 
features that deter seed predators. The most often cited mechanism is the irregularity of seed crops from 
year to year. Most nut-producing trees (oak, hickory, walnut, beech) flower in the early spring. This 
allows wind to transport pollen through the bare canopy, and causes late frost and spring storms to 
damage seed crops in certain years. Ecologists suggest that the natural cycling of the seed crop serves to 
keep predator populations under control. Nuts are also protected by thick, hard seed coats and/or 
chemicals such as tannins. 

Chestnut is unique among nut producing trees in North America in having essentially no deterrence for 
seed predators. The seed itself is protected by a thin shell and contains no tannins or other chemicals to 
reduce palatabilit>'. Young chestnuts are protected by a formidable burr during development, but the burr 
opens wide when the seed is ripe. The maturing fruit is exposed to possible damage for a relatively short 
period and is protected from predation by its burr during that period. Chestnut flowers in late June or July 
when there is no possibility of frost. All of these factors show that chestnut fruit is produced with regular 
seed crops that are designed to be as attractive to predators as possible. This presents a paradox in that 
chestnut seems to defy the logic of predator/prey cycles. 



47 



Various studies suggest that seed predation is a real problem in chestnut reproduction. All of the early 
forestry literature cites a lack of chestnut reproduction by seed. Buttrick and Holmes (1913) suggest that 
lack of seedling establishment may be the cause of the loss of chestnut from some North Carolina 
piedmont sites where it was formerly abundant. Thoreau (1906) likewise describes a nearly complete 
lack of chestnut seedlings in chestnut woodlots in Massachusetts. Paillet (1988) noted that there were 
some New England sites where logs and stumps showed that chestnut was once dominant in the canopy, 
but where the low density or lack of living sprouts shows that chestnut seedlings were not being 
established. He also noted that maps of living chestnut sprouts indicated microsites such as fences, rights 
of way and brush thickets influenced seedling establishment. Pridnya et al (1996) note a similar condition 
in old-growth European chestnut forests in southern Russia. Since it is unlikely that these forests were 
not producing chestnuts and since there is no obvious reason why some seedlings would not germinate, 
seed predation by livestock or wildlife is assumed to have caused the lack of surviving seedlings. Such 
observations leave no doubt that seed predation can be a significant problem for chestnut reproduction. 
The nearly complete lack of any deterrence with regard to seed predation and the apparent effects on seed 
predation for at least the short term in many historic woodlands presents an ecological parado.x that ma\ 
have implications in attempts to restore naturally-reproducing chestnut trees to American forest. 



A POSSIBLE EXPLANATION 

My hypothesis addresses this paradox under the assumption that chestnut is a successful forest tree and 
recognizable as a distinct genus in the fossil record for at least 50 million years (Graham. 1990). If 
chestnut is a successful species, why has competitive interaction with other forest trees not selected seed 
traits for factors that deter seed predation? The simplest answer is that such adaptations have not arisen 
because there is no significant adaptive advantage to them. I suggest that the unique seedling re-sprouting 
capability of chestnut produces "immortal" seedlings that can sur\ ive for centuries. 1 propose that the 
longevity of viable seedlings (viable in the sense that they can be released to form a cleanly-formed, 
canopy-dominant stem) completely compensates for seed predation. At the same time, the extended life 
cycle of chestnut as a tree (centuries in the understory and then several centuries in the canopy) implies 
that chestnut "invasion" of new territor> is a slow process. This, in turn, places added significance on 
mechanisms for seed dispersal. I conclude that seed predation is not important for chestnut because the 
longevity of a few established seedlings compensates for predation. At the same time, the inherent 
slowness in forest turnover associated with this process places a premium on seed dispersal. Chestnut is 
"designed" to encourage seed transport in every way, and is largely unaffected by seed predation in the 
long tenn. 

This explanation is largely a "default option" that suggests chestnut does not deter seed predation because 
such deterrence has a metabolic cost but has no selective advantage. Is there an\ positive support for this 
idea? One line of evidence comes from the comparison of chestnut pollen on two adjacent sites in 
Massachusetts (Fostei and Zybruk, 1991 ). The data give profiles for a bog catchment receiving pollen 
for many square km around the site, and a forest hollow catchment underneath the forest canop\ receiving 
mostly local pollen from directly abo\e. The bog pollen shows a rather consistent regional proportion of 
10% chestnut pollen over the past 3000 years. In contrast, the forest hollow pollen shows lluctuations 
from near zero to 60% or more that follow discrete disturbance events (fire and w indstonn). This 
suggests that the regional proportion of chestnut was constant, but that the local abundance of chestnut 
was highly variable. I interpret this as a situation where seed predation concentrated on the areas 
immediate around and beneath mature trees, while chestnut seedlings could escape predation in other 
parts of the forest. I'his is exacll\ the same situation as reported b\ Thoreau ( 1906). where he could not 
find chestnut seedlings in chestnut woodlots. but found them ahundanl in adjacent old-tlcld pine stands. 
Russian ecologists identify a similar "bottleneck" in chestnut reproduction related to the need to establish 
chestnut seedlings in the understory (Pridnya et al. 1996; figure 5) 



48 



Does any of this have a bearing on the introduction of blight resistant chestnut into national park forests? 
This hypothesis explains why chestntit is such a slowly migrating species under the influence of climate 
change. Introduction of blight-resistant chestnut would also be a lengthy process. Several approaches 
coulo be used to hasten the process. First, one could use special seed-predator exclusion techniques to 
produce microsites suitable for chestnut seedling establishment. One could also transplant seedlings to 
generate established old seedlings to bypass the reproductive "bottleneck" related to seedling 
establishment. Then one could also rely on the artificial generation of disturbance to promote the release 
of suppressed seedlings. Although such manipulation may seem ill suited for wilderness areas in national 
parks, the alternative is a period of 1000 years or more before restored chestnut reaches a natural 
equilibrium with the surrounding forest. 



SUMMARY 

The interaction of chestnut with wildlife can be separated into two issues: 1 ) Chestnuts as a resource for 
wildlife consumption; and 2) Wildlife as a dispersal agent for chestnut seed. Making an analogy between 
the growth form and nut crop of northern red oak and chestnut, the former chestnut mast resource is 
estimated to be qualitatively similar to that of other large-fruited mast species. Chestnut mast may have 
been quantitatively better because chestnut crops were probably more reliable from season to season, and 
because chestnut has been partly replaced by tree species such as tulip poplar that do not produce nuts. 
These factors may result in a 20 to 50% increase in total mast production if chestnut is re-established as a 
canopy-dominant tree in American forests. The availability of chestnuts in years of frost or insect 
damage to acorns may make chestnut an especially valuable addition to the forest. Dispersal of chestnut 
seed and seed burial by wildlife was important because chestnuts are relatively fragile and have little 
resistance to frost or desiccation. However, chestnuts as seed exhibit none of the mechanisms Msually 
associated with a trade-off between dispersal and predation. Chestnut reproduction occurred most often 
by coppice sprouting, and chestnut seedlings were reportedly rare or absent in many chestnut stands. One 
explanation for this apparent paradox is that there is no selective pressure on chestnut for the deterrence 
of predators. This happens because chestnut reproduction is based on a combination of wildlife dispersal 
of nuts and long-term survival of the few seedlings that do manage to germinate. Many living chestnut 
seedlings have survived for at least a century as small suppressed stems and possess root collar sprouting 
characteristics designed to insure that "old seedlings" remain juvenile essentially forever. Thus, chestnut 
reproduction by seed is heavily biased in favor of mechanisms to promote long-distance transport. The 
near immortality of established chestnut rootstocks could offset any selective value in deterring seed 
predation within chestnut-dominated stands. 



49 




Figure 5. Many European chestnut (Castanea sativa) trees in old growth forests of southern Russia show 
multiple stems indicative of origin as sprouts derived from suppressed seedlings (from Pridnya et al, 
1996). 



LITERATURE CITED 

Adams. S.M., and S.L. Stephenson. 1983. A description of the vegetation on the south slopes of Peters 
Mountain, southwestern Virginia. Bull. Torrey Bot. Club 1 10:18-23. 

Anderson, T.W. 197^. The chestnut pollen decline as a time horizon in lake sediments in eastern North 
America. Can. J. Earth Sci. 1 1 :678-685. 

Boring, L.R., CD. Monk, and W.T. Swank. 1981. Early regeneration of a clear-cut southern Appalachian 
forest. Ecology 62:1244-1253. 

Brugham, R.B. 1978. Pollen indicators of land-use change in southern Connecticut. Quaternary Res. 
9:349-362. 

Buttrick. P.L., and J.S. Holmes. 1913. Preliminaiy report on the chestnut in North Carolina made in 
connection with a cooperative investigation of the chestnut bark disease. North Carolina Geological and 
Economic Survey, Raleigh, NC. 



50 



Davis, M.B. 1969. Climate changes in southern Connecticut recorded by pollen changes at Rogers Lake. 
Ecology 50:409-522. 

Davis. M.B., R.W. Spear, and L.C.K. Shane. 1980. Holocene climate of New England. Quaternary Res. 

14:240-250. 

Foster, D.R., and T.M. Zebryk. 1991. Long-temi vegetation dynamics and disturbance history of a Tsuga- 
dominated forest in central New England. Ecology 74:982-998. 

Frothingham, E.H. 1912. Second growth hardwoods in Connecticut, USDA Bull. No. 95. 

Graham, A. 1990. Late Cretaceous and Cenozoic history of North American vegetation. Oxford 
University Press, NY. 350 p. 

Hibbs, D.E. 1983. Forty years of forest succession in central New England. Ecology 64:772-783. 

Krochmal, A., and C. Krochmal. 1982. Uncultivated nuts of the United States. Ag. Infonnation Bull. 450, 
USDA Forest Service, Washington, DC. 89 p. 

Matoon, F.E. 1909. The origin and early development of chestnut sprouts. Forest Quarterly 7:34-37. 

Nichols, G.E. 1913. The vegetation of Connecticut. Torreya 13:19-1 12. 

Paillet, F.L. 1984. Growth fonn and ecology of American chestnut sprout clones in northeastern 
Massachusetts. Bull. Torrey Bot. Club 1 1 1:316-328. 

Paillet, F.L. 1988. Character and distribution of American chestnut sprouts in southern New England 
woodlands. Bull. Torrey Bot. Club 1 15:32-44. 

Paillet, F.L. 1993. Growth fonn and ecology and life history of American chestnut and Allegheny 
chinquapin at various North American sites. Bull. Torrey Bot. Club 120:257-268. 

Paillet, F.L. 2003. Chestnut: history and ecology of a transformed species. Biogeography 29:1517-1530. 

Paillet, F.L., and P. A. Rutter. 1989. Replacement of native oak and hickory tree species by the introduced 
American chestnut {Castcmea dentata) in southwestern Wisconsin. Can. J. Bot. 67:3457-3469. 

Paillet, F.L., M.G. Winkler, and P.R. Sanford. 1991. Relationship between pollen frequency in moss 
polsters and forest composition in a naturalized stand of American chestnut: implications for 
paleoenvironmental interpretation. Bull. Torrey Bot. Club 1 18:432-443. 

Pridnya, M.V., V.V. Cherpakov, and F.L. Paillet. 1996. Ecology and pathology of European chestnut 
{Castauea sativa) in the deciduous forests of the Caucasian mountains of southern Russia. Bull. Torrey 
Bot. Club 123:213-222. 

Russell, E.W.B. 1987. Pre-blight distribution of Castauea dentate. Bull. Torrey Bot. Club 1 14:180-193. 

Saucier, J.R. 1973. American chestnut - an American wood. USDA For. Serv. Rep. FS-230. 



51 



Smith, D.M. 2000. American chestnut- ill-fated monarch of the eastern hardwood forest. J. For. 98:12- 

15. - - 

Stephens, G.R., and P.E. Waggoner. 1980. A half century of natural transition in a mixed hardwood 
forest. Conn. Agr. Exp. Sta. Bull. 783. 

Stephenson, S.L., H.S. Adams, and M. Lipford. 1991. The present distribution of chestnut in the upland 
forest communities of Virginia. Bull. Torrey Bot. Club 1 18:24-32. 

Thoreau, H.D. 1906. Journal, Vol. XIV. 

Watts. W.A. 1979. Late Quaternany vegetation of the central Appalachians and the New Jersey coastal 
plain. Ecol. Monogr. 49:427-469. 

Whitehead, D.R. 1979. Late glacial and postglacial vegetational history of the Berkshires. Quaternary 
Res. 12:333-357. 

Zon, R. 1904. Chestnut in southern Maryland, USDA Bureau of Forestry, Bull. No. 53. 



52 



Steiner. K. C. and Carlson, J. E, eds. 2006. Restoration of American Chestnut To Forest Lands - Proceedings of a 
Conference and Wori<shop. May 4-6, 2004, The North Carolina Arboretum. Natural Resources Report 
NPS/NCR/CUE/NRR - 2006/001, National Park Service. Washington, DC. 



HISTORICAL SIGNIFICANCE OF AMERICAN CHESTNUT 

x^ ' TO APPALACHIAN CULTURE AND ECOLOGY 

Donald E. Davis 
Social Sciences Division. Dalton State College, Dalton, GA 30720 USA (ddavis@em.daltonstate.edu) 



Abstract: This paper explores the significance of the American chestnut on the ecology and culture of 
Appalachia. Until the third decade of the 20"' century, the tree was the crowning glory of the Appalachian 
hardwood forest, in some isolated areas comprising half of the hardwood tree population. The wildlife of 
the region, particularly black bears, heavily depended on the tree for both sustenance and shelter. Native 
Americans in the mountains frequently made use of the nut, mixing chestnut meal with corn to make 
bread. White mountaineers gathered the nuts to sell or trade, and sometimes used parched chestnuts as a 
coffee substitute. The American chestnut played a major role in the economy of the Appalachian region, 
providing timber for dwellings and tannic acid for the leather industry. Finally, it is argued that the 
decline of Appalachian subsistence culture is directly linked to the loss of the American chestnut. 



INTRODUCTION 

Few single events in North American environmental history compare with the loss of the American 
chestnut. "The devastation of the American chestnut by the chestnut blight," wrote William MacDonald 
more than two decades ago. "represents one of the greatest recorded changes in natural plant population 
caused by an introduced organism." MacDonald, a professor of plant pathology at West Virginia 
University and the current acting treasurer of the American Chestnut Foundation, estimates that 
chestnut-dominated forests once covered 200 million acres of land from Maine to Mississippi 
(MacDonald 1978; Brown and Davis 1995). American chestnut trees once comprised roughly twenty 
percent of the Appalachian forest, although in specific areas they accounted for as much as one-third of 
all trees (Kulman 1978). In fact. William Ashe reported seeing several locales in western North Carolina 
where the trees "occur pure or nearly pure over areas as large as 100 acres" (Ashe 1912). In 1901, he and 
Horace Ayers estimated that the southern Appalachians contained more than 884.000 acres of chestnut 
timber (Ayres and Ashe 1905). 

The American chestnut was confined largely to the Blue Ridge Mountains and Cumberland Plateau where 
the trees commonly grew at altitudes between 1 .000 and 4.000 feet. According to Charlotte Pyle. the 
American chestnut covered no less than 3 1 percent, or 1 59,165 acres, of what is today the Great Smoky 
Mountains National Park (Pyle 1985; Davis 2000). The Ridge and Valley physiographic province had a 
few important stands of chestnuts as well, but these were found only on the slopes of the highest ridges 
where richer soils and heavier rainfall predominated. A notable exception was a nearly one-square mile 
area of Walker County. Georgia, still known today as Chestnut Flats. According to historian James 
Sartain. the area "was so-called because of the abundance of chestnuts that grew in that beautiful valley 
and on the adjacent ridge when the early settlers arrived" (Sartain 1 972). A reconstruction of nineteenth 
century forests in other parts of northwest Georgia, however, found chestnut trees comprising no more 
than six percent of the area, with oaks and hickories, the most dominant tree species, making up 45 
percent of the total forest (Plummer 1975). 



53 



As is now well-documented, the death of the American Chestnut was due to an exotic blight introduced in 
the United States from Japanese Chestnut nurser> stock just after the turn of the centurv. A forester at the 
New York Zoological Park first reported the disease in 1904, after observing an immense number of dead 
and dying chestnut trees on park lands under his supervision. Five years later, the first scientific bulletin 
appeared about the disease, a fungus later named Endothia parasitica (Murrill 1908; Gravatt 1930; 
Hepting 1974; Kulman 1978). Only a year after the bulletin's publication, an editorial in the Southern 
Lunihernuin referred to a "mysterious blight" that had recently been observed in Penns\ Ivania and New 
York. "Large timbered sections of [Pennsylvania] are already and in an alarming manner affected by the 
disease," stated the report (Southern Lumberman 1910). By 1912 all the chestnut trees in New York City 
were dead and the chestnut blight had reached no fewer than 10 states. Scientists in Pennsylvania 
launched a vigorous control program, which included burning dead trees, monitoring its advance, and 
spraying infected trees (Anagnostakis and Hillman 1992). This effort, a scientist later commented, was a 
little like using toy swords to battle an enemy equipped with atomic bombs. At the same time, foresters 
told the public that "the control and ultimate e.xtemiination of [the chestnut blight] will sooner or later 
become a real accomplishment" (Brown and Davis 1995). 

The disease spread relentlessly southward, at an astounding rate of some fifty miles per year. Aided by 
woodsmen and loggers who carried it on their shoes and axes, the blight first entered North Carolina near 
Stokes and Surry counties about 1913 (Buttrick 1925). Shady Valley, in upper east Tennessee, was hit by 
1915 (Cole 1990). By 1920 the American chestnut in the Great Smoky Mountains was ultimately 
doomed, though there were few visible signs of the blight there before 1925 (Brown 2001). In nearby 
Yancey County, North Carolina, nearly one in every ten chestnut trees was showing signs of the disease 
by 1925; in Buncombe County, one in five trees was dying from the blight at that time (Silver 2003). 
North Carolina lumbermen even used the imminentiv encroaching disease as a last-ditch effort to defeat 
the proposed Great Smoky Mountains National Park. "Certainly nothing could be more unsightK than 
the gaunt and naked trunks of these dead trees, standing like skeletons in every vista which the e\e turns," 
they wrote in 1931 (Baxter 1931). By the mid- 1930s, the blight had reached north Georgia, and by 1940 
there was scarcely a tree in the entire Appalachian region that was not dead or showed signs of being 
severely infected with the disease (Exum 1992; Davis 2000). 



CHESTNUT MEMORIES 

Although few people alive today remember what the Appalachian forests looked like before the blight 
devastated the region, those who did witness the trees in their native splendor provide indisputable 
testimony to their significance to the mountain environment. "This is an unbelievable thing: how many 
chestnuts there were," remembered Paul Woody, who grew up near Cataloochee, North Carolina (Wood\ 
1973). Gifford Pinchot himself recalled seeing chestnut stands with individual trees thirteen feet across 
and with crowns spreading more than 120 feet above the forest floor (Wheeler 1988; Davis 2000). 
Writing in the October 1915 issue of American Forestry, Samuel Detvviler noted that the "finest chestnut 
trees in the world are found in the southern Appalachian Mountains." adding that a tree w ith a diameter of 
seventeen feet had been found in Francis Cove, North Carolina (Detwiler 1915). Charles Grossman, one 
of the first rangers at the new Great Smoky Mountains National Park, recorded a chestnut tree 9 feet, 8 
inches in diameter at a point six feet off the ground. "The hollow portion is so large that [an adult] could 
stand up in it." wrote Grossman soon after discovering it. "This hollow runs more than 50 feet up the 
trunk and at its narrowest point is not less than three feet. This must be the tree of which 1 heard. A man 
lost some stock during a snowstorm and later found (hem safe in a hollow chestnut tree" (\\ heeler 1988; 
Davis 2000; Brown 2001). 

Due to their abundance and enomious size, the American chestnut ranked as the most important w lid life 
plant in the eastern United States. The largest trees could produce ten bushels or more of nuts. Reports of 



54 



chestnuts four inches deep on the forest floor were not uncommon in many parts of the Appalachian 
mountains. Many of the wildlife species that mountain people thought of as game--squirrels, wild turkey, 
white-tailed deer, black bear, raccoon and grouse-depended on these chestnuts as a major food source. 
"] he worst thing that ever happened in this country was when the chestnut trees died." recalled Walter 
Cole of east Tennessee. "Turkeys disappeared, and the squirrels were not one-tenth as many as there 
were before... bears got fat on chestnuts, coons got fat on chestnuts, and the woods was filled with wild 
turkey. ..most all game ate chestnut..." (Cole 1965). Will Effler, who grew up on the West Fork of the 
Little River in what is today the Great Smoky Mountains, recalled shooting a wild turkey that contained 
no fewer than ninety-two chestnuts, "still in the hulls and undigested" in its swollen craw (Weals 1991 ). 
The former Cades Cove resident Maynard Ledbetter once remarked that "back when there were chestnuts, 
bear got so fat they couldn't run fast; now the poor bear run like a fox" (Ledbetter 1989). 

Non-game animals were equally dependent on the chestnut, including several unique insect species that 
relied upon chestnut trees as their principal food course. Paul Opler, formerly of the U.S. Fish and 
Wildlife Service, has estimated that at least seven native moths became extinct in the southern 
Appalachians as a result of the chestnut blight (Opler 1978). The chestnut also slowed the recovery of 
wildlife populations already suffering from loss of habitat by logging operations. Biologist James M. Hill 
ascribes the slow recovery of deer, wild turkey, goshawks. Cooper's hawks, cougar, and bobcat in the 
mountains to habitat destruction directly caused by the chestnut blight (Hill 1993). 

Of course, humans seasonally ate chestnuts too. making them an important dietary supplement when the 
trees dropped their nuts after the first major frost. Each October, children living in the mountains scooped 
up chestnuts by the sackful, often hanging their cloth bags on nails outside the kitchen door until 
December when the nuts would begin to get wormy. Smoky Mountain resident Alie Newman Maples 
remembered: "As a little girl, me and my brother Ray would take a sack or a pail and go out to the woods. 
Strong winds blew in the night, and we would pick up gallons of chestnuts under each tree" (Miples 
1973; Brown 2001 ). Environmental historian Margaret Brown notes in her book Wild East: A Biography 
of the Great Smoky Mountains, that many mountain families routinely baked chestnuts in the kitchen 
fireplace, roasting them in dutch ovens. Among her most notable entries are the chestnut memories of 
Delce Mae Carver, who remembered sackfuls of chestnuts hanging on nails near the kitchen, ready to be 
baked over a warming fire. Johnny Manning, another Smoky Mountain resident who grew up in 
Greenbrier Cove, recalls as a child "trading pocketfuls of chestnuts for school tablets and pencils" (Brown 
2001). For some Smoky mountain residents, the earnings from fall chestnut gathering was known as 
"shoe money," as the funds were used to purchase children's shoes before the coming winter (Brown and 
Davis 1992; Condon 1994). 

Cherokees in Appalachia made even more use of the nut, which they frequently added to cornmeal dough 
that "was boiled or baked." Cherokees also used leaves from the tree to alleviate heart troubles, and the 
sprouts were sometimes made into an astringent tea to treat healing sores and wounds (Wigginton 1972; 
Stewart in press). All mountain families gathered many bushels of chestnuts, often taking them by wagon 
to urban markets. John McCaulley, whose family foraged for chestnuts in the Great Smoky Mountains 
around 1910. remembered seeing in one mountain cabin, a "hundred bushels of chestnuts, piled up there, 
and about four men packing off every day." McCaulley himself recalls gathering as many as seven 
bushels of chestnuts in a single day's outing. These, he said, were taken to Knoxville on mules where 
they were sold for "four dollars a bushel" (Brown and Davis 1992). Chestnuts were also routinely 
shipped by rail to major cities on the eastern seaboard. In 1911, West Virginia reported that one railroad 
station alone shipped 155,000 lbs. of chestnuts to destinations along the train's northerly route (Giddings 
1912;Kulman 1978). 

Another historical use of chestnuts in the mountain region was food for hogs. Frederick Law Olmstead, 
in his travels through the Appalachians in 1854, reported that raising hogs was "remarkable fine" in the 



55 



mountains due to the large chestnut mast crop. He also noted that the swine of the region were of 
''superior taste" than those raised elsewhere in the South, a fact that made mountain pork a much sought 
after commodity (Olmsted 1860; Weals 1981). fhe huge annual mast production made woodland grazing 
possible, so for a month or two each fall, hogs ran loose in the woods to feast on the chestnuts littering the 
forest floor. Martha Wachacha, recalling the scene around her home in Cherokee. North Carolina, said 
"there were about a hundred pigs when I first moved here. Pigs and hogs were so fat. There was plenty of 
chestnuts back then" (Wachacha 1989). In late November, or as soon as the weather got cold enough, 
mountain residents rounded up the fattened hogs for slaughter. Martin Tipton recalled that "mountain 
people needed those chestnuts. They ate them themselves, of course, but they depended upon them to 
feed their hogs" (Brown and Davis 1994). Chestnut-flavored pork hung in the smokehouse all winter, 
where it continued to be the primary source of protein for most families. A Virginia farmer commenting 
on the role of chestnuts in mountain agriculture noted that it "didn't cost a cent to raise chestnuts or hogs 
in those days. It was a very inexpensive way to farm. The people had money and had meat on the table 
too" (Nash 1988). 

As a building material, chestnut timber was unsurpassed. Chestnut wood was also highly rot-resistant, 
making it ideal for roofing shingles, telephone poles, ship masts, railroad ties and almost any other use 
requiring durable, long-lasting timbers. In 1909, the timber industry placed the total value of chestnut 
timber in the United States at more than 20 million dollars (Stewart 2005). Builders found chestnut wood 
to be remarkably insect-proof and weather resistant, so chestnut logs made the best fence rails, fence 
posts, and caskets. "Chestnut wood," as George Kulman wryly noted, "carried man from cradle to grave, 
in crib and coffin" (Kulman 1978). Seymour Calhoun, a full-blood Cherokee, added that "it was soft 
wood and worked good: you could split it" (Calhoun 1973). Chestnut trees grew so large that in one 
documented case, an entire cabin in the Great Smoky Mountains National Park was constructed from a 
single tree (Brown 2001 ). A valuable source of tannic acid used in the leather industrv. chestnut bark and 
rough chestnut cordwood was another important source of income for mountain residents. In Tennessee 
alone, 50,000 cords of wood were cut yearly to supply those tanneries in operation before 1912. This 
"tanbark" or "acid wood," as it was called locally, was taken largely from trees already cut for other 
purposes or small defective trees that were not of nut-bearing age. Commercial operations were also 
heavily engaged in the harvesting of chestnut trees for tanbark and cordwood. One observer remarked in 
193 1 that even though chestnut timber was once cut by lumbermen for the bark alone, "ver> little waste of 
this kind is now noted" (Frothingham 1925). 

As might be expected during the era of industrial logging, the blight did not slow the har\est of chestnut 
trees; in fact, the cutting actually increased after the initial introduction of the disease. In fact, most 
lumber barons were harvesting the largest chestnut trees even before the blight was officially obser\ed in 
the mountain region. Early on. lumbemien even doubted the potential devastation of the disease, 
believing that the fast-going trees would eventually regenerate across the mountain landscape. Moreover, 
they knew that a chestnut tree was worth money dead or alive, since foresters soon determined that it was 
possible to manufacture lumber from standing dead chestnuts for up to ten years after the death of the 
tree. In fact, "wormy chestnut" lumber became much sought after bv builders and furniture makers alike 
for many decades to come. For acid wood, the salvage period was even longer: Reuben Robertson, then 
president of the Champion Fibre, estimated that the companx cut chestnut trees for pulp and tannin twenty 
years after the blight first arrived in North Carolina (Nelson and Gravitt 1929; Robertson 1959). 



A WHOI.F WORLD DYING 

The abundance of dying chestnut trees was also responsible for the expansion and growth of the region's 
leather tanning industry. By 1930, there were no fewer than twenty-one chestnut-fueled plants in the 
southern Appalachians, producing over one-half of the U.S. supply of vegetable based tannins. Within a 



56 



decade, however, almost all evidence of chestnut trees had vanished from the mountains as the growing 
tanning industry, the "largest consumer of chestnut," had found ways to use every part of the tree. After 
1940, with the development of synthetic replacements in the production of tannin, the demand for 
chestnut greatly diminished, leaving only a few ghost-white skeletons to stand lone sentry over the once 
great Appalachian forest. The dead and dying chestnut snags were painful reminders to mountaineers that 
the mountain landscape, including an entire way of life, was all but gone. "Man. I had the awfulest 
feeling about that as a child, to look back yonder and see those trees dying." recalled Joe Tribble, a native 
of eastern Kentucky. "1 thought the whole world was going to die" (Hawkings 1993). A similar 
sentiment was echoed by Martin Tipton, who remembered that he and his dad used to come upon the 
skeletons of the trees on their many mountain walks. "Dad said it looked like a third of the mountain was 
dying" recalled Tipton (Brown and Davis 1994). 

Mountain residents were right to mourn the lost of the American chestnut. The chestnut tree was possibly 
the single most importance natural resource of the Appalachians, providing inhabitants with food, shelter, 
and in the early twentieth century, a much needed cash income. Knott County. Kentucky native Verna 
Mae Sloan recalled that life without the chestnut tree was almost unthinkable. "At first we thought they 
would come back, we didn't know they were blighted out forever." she remembered. "But the chestnut 
tree was the most important tree we had. We needed those chestnuts" (Sloan, Pers. Comm., 1998). In fall 
and winter chestnuts could be boiled or roasted over an open fire or traded at the local stores for much 
needed supplies. Having "the greatest durability of available native woods." chestnut timber was made 
into long-lasting boards, posts, shingles, and split-rail fences. The tender and abundant sprouts could 
even be pulled from the ground and fed to cattle as fodder. As a wildlife food, the chestnut was 
unsurpassed, and helped to keep local game populations at their highest levels in recorded memory. In a 
memoir written shortly before his death. Shady Valley, Tennessee native William Cole summed up the 
extraordinary value of the tree to mountain residents. "A favorite outing for me and my friends was to go 
to the ball ground on Sunday to collect chestnuts," wrote Cole. "The chestnut tree was a great t'-ee. 
chestnut wood was a great wood, and chestnuts a good food" (Cole 1990). 

Sadly, the chestnut blight made it very unlikely that the Appalachian mountaineers would return to their 
more self-sufficient way of life. By the late 1930s, the mountaineer was more off the famistead than on 
it. as the food and folkways of the region's inhabitants were beginning to conspicuously change. By the 
early 1940's mountain families were utilizing less buttermilk and more whole milk, less r>'e and wheat 
breads and more light breads, and consuming more processed sugar and less maple syrup and honey. 
While their were some dietary constants throughout the regioi. such as the consumption of combread and 
biscuits, the use of canned and other "store-bought" foods increased significantly during the first three 
decades of the twentieth century (Wheeler 1935). For those who remained exclusively farmers, the 
practice of crop monoculture became a much more common way to farm. Family size dropped by more 
than two individuals, from 10 family members per household in 1910 to 7.62 per household in 1934. 
Home building techniques changed as well. "Boxed" houses-that is. frameless structures made 
exclusively with sawn planks and boards-gradual ly replaced log cabins as residents working seasonally 
for lumber companies had less time, or the extra help, to build traditional log homes. The number of 
working outbuildings on the homestead also diminished, including the smokehouse, spnnghouse. and 
separate kitchen facility. Furniture was no longer home-made and looms and spinning wheels largely 
became a thing of the past (Black 1928). Needless to say. everything from architecture to social relations 
was altered by the separation of the mountain environment from the mountaineer. 

In many ways, the death of the American chestnut symbolized the end of a waning, albeit arguably vital, 
subsistence culture in the Appalachians. The loss of the tree no doubt gave additional advantage to the 
forces of industrialization that were gaining a stronger and stronger foothold on the regional and local 
economy. No longer able to range hogs and cattle in the woodland commons, trap fish in free-fiowing 
streams, or gather chestnuts on the hillsides, the rural mountaineer increasingly looked to the milltown 



57 



and urban center for economic salvation. The environmental abuse of the mountains, along with their 
permanent removal from the traditional land base, made it extremely difficult for mountaineers to 
continue a semi-agrarian, and intimately forest-dependent, way of life. With the death of the chestnut, an 
entire world did die, eliminating subsistence practices that had been viable in the Appalachian Mountains 
for more than four centuries. 



LITERATURE CITED 
Anagnostakis, S.L., and B. Hillman. 1992. Evolution of the chestnut tree and its blight. Arnoldia 52:3-10. 
Ashe, W.W. 1912. Chestnut in Tennessee. Tenn. Geological Survey Bull. 10-B. 35 p. 

Ayers, H.B., and W.W. Ashe. 1905. The southern Appalachian forests. U.S. Geological Sur\'e\ Prof Pap. 

37. 232 p. 

Baxter, D.V. 1931. Deterioration of chestnut in the southern Appalachians. USDA Technical Bull. 257. 
22 p. . 

Black, E.E. 1928. A study of the diffusion of culture in a relatively isolated mountain community. Ph.D. 
dissertation, University of Chicago, Chicago, Illinois. 134 p. 

Brown, M., and D. Davis. 1992. Trail history notebook. Great Smoky Mountains Natural History Assoc, 
Gatlinburg, Tennessee. 182 p. 

Brown, M., and D. Davis. 1994. Old Settlers Trail. P. 439-446 in Hiking trails of the Smokies, DeFoe, D., 
et al. (eds.). Great Smoky Mountains Natural History Assoc, Gatlinburg, Tennessee. 

Brown, M., and D. Davis. 1995. 1 thought the whole world was going to die. Now and Then 12:30-31. 

Brown, M.L. 2001 . Wild east: A biography of the Great Smoky Mountains. University of Florida Press, 
Gainesville, Florida. 457 p. 

Buttrick, P.L. 1925. Chestnut in North Carolina. North Carolina Geological and Economic Survey 
Economic Pap. 56. 10 p. 

Calhoun, S. 1973. Interview by William F. Alston. Transcript in Oral llistorv collection. Great Smoky 
Mountains National Park. .A.rchives, Sugarlands Visitor Center, Gatlinburg, Tennessee. 

Cole, W. 1965. Intei'view by Charles Grossman. Transcript in Oral Historv collection. Great Smoky 
Mountains National Park Archives, Sugarlands Visitor Center. Gatlinburg. Tennessee. 

Cole, W.E. 1990. Tales from a country ledger. Tapestry Press, Acton. MA. 

Condon. T. 1994. Chestnut top trail. P. 166-168 in Hiking trails of the Smokies. DeFoe. D., et al. (eds.). 
Great Smoky Mountains Natural Historv' Assoc. Gatlinburg. Tennes.see. 

Davis. D.E. 2000. Where (here are mountains: An en\ ironmcntal histoi"\ of the southern Appalachians. 
University of Georgia Press, Athens, GA. 320 p. 

Detwiler, S.B. 1915. The American chestnut tree. Am. Forestry 2l(262):957-960. 



58 



Exum, E.M. 1992. Tree in a coma. Am. Forests 28(1 1/12):20-26. 

Frothingham, E.H. 1925. The present stand of chestnut in North Carolina and in the southern 
Appalachians. Geological and Economy Survey Economic Pap. 56. 7 p. 

Giddens, N.J. 1912. Untitled report on chestnut blight. P. 173-174 in Proc. of conf. on Chestnut blight. 
Harrisburg, Pennsylvania, February 20-21. 

Gravatt, G.F. 1930. Chestnut blight. USDA Dept. of Ag. Farmers' Bull. 1641. 3 p. 

Hepting, H.G. 1974. Death of the American chestnut. J. For. Hist. 18:60-67. 

Hawkings, N. 1993. Building community through grassroots democracy. Local Voices 10(2/3):5-8. 

Hill, J.M. 1993. Wildlife value oiCastcmea detilata past and present, the historical decline of the 
chestnut, and its future use in restoration of natural areas. Unpublished manuscript, Randolph Macon 
College, Lynchburg, Virginia. 

Kuhlman, E.G. 1978. The devastation of American chestnut by blight. P. 1-3 in Proc. of the American 
chestnut symposium, MacDonald, W.L., et al. (eds). West Virginia University Press, Morgantown, WV. 

Ledbetter, M. 1989. Interview by Bill Landry. Transcript in Landry Collection, Great Smoky Mountains 
National Park Archives, Sugarlands Visitor Center, Gatlinburg, Tennessee. 

Maples, A.N. 1973. Interview by Jane Whitney. Trancript in Oral History Collection, Great Smoky 
Mountains National Park, Sugarlands Visitor Center, Gatlinburg, Tennessee. 

MacDonald, W.L. 1978. Foreward. P. v in Proc. of the American chestnut symposium, MacDonald, W.L., 
et al. (eds). West Virginia University Press, Morgantown, WV. 

Metcalf, H. 1910. The chestnut tree blight: An incurable disease that has destroyed dollars worth of trees. 
Sci. Am. 106:241-42. 

Murrill, W.A. 1908. The spread of the chestnut disease. J. N.Y. Bot. Gard. 9:23-30. 

Nash, S. The blighted chestnut. National Parks 62:14-19. 

Nelson, R.M., and G.F. Gravatt. 1929. The tannin content of dead chestnut trees. J. Am. Leather Chem. 
Assoc. 24:479-99. 

Olmsted, F.L. 1860. A journey in the backcountry, 1853-1854. Ben Franklin. New York, NY. 

Opier, A. P. 1978. Insects of American chestnut: possible importance and conservation concern. P. 83-85 
in Proc. of the American chestnut symposium, MacDonald, W.L., et al. (eds). West Virginia University 
Press, Morgantown, WV. 

Plummer, G.L. 1975. 18"" century forests in Georgia. Bull. Geor. Acad. Sci. 33:1-19. 

Pyle, C. 1985. Vegetation disturbance history of the Great Smoky Mountains National Park. Unpublished 
manuscript. Uplands Laboratory, Gatlinburg, Tennessee. 



59 



Robertson, R. 1959. Interview by Jerrv Mander. Vertical Files, Great Smoky Mountains National Park 
Archives, Sugarlands Visitor Center, Gatlinburg, Tennessee. 



Sartain, J. A. 1972. History of Walker County, Georgia. Thomasson Printing & Office Eqpt. Co, 
Carrollton, Georgia. 559 p. , 

Silver, T. 2003. Mount Mitchell and the Black Mountains: An environmental history of the highest peaks 
in eastern America. University of North Carolina Press, Chapel Hill, NC. 322 p. 

Southern Lumberman. 1910. Editorial. Southern Lumberman. 1 10:38C. ^-_ 

Stewart, C.J. (in press) The American chestnut blight. The Encyclopedia of Appalachia, Abramson. R.. et 
al. (eds.). University of Tennessee Press, Knoxville, TN. 

Weals, V. 1991. Last train to Elkmont. Olden Press, Knoxville, TN. 150 p. 

Wheeler, D. 1988. Where there be mountains, there be chestnuts. Katuah J. 21(3):3-5. 

Wheeler, L.R. 1935. Changes in the dietary habits of remote mountain people since 1900. J. Tenn. Acad. 
Sci. 10:167-74. 

Wiggington, E (ed.). 1972. The foxfire book. Doubleday & Company, Garden City, NY. 384 p. 

Woody, P. 1973. Interview by Katherine Manscill. Transcript in Oral Histor\ Collection. Great Smoky 
Mountains National Park Archives, Sugarlands Visitor Center, Gatlinburg, Tennessee. 



60 



Steiner, K. C. and Carlson, J. E, eds. 2006. Restoration of American Chestnut To Forest Lands - Proceedings of a 
Conference and Workshop. May 4-6, 2004, The North Carolina Arboretum. Natural Resources Report 
NPS/^CR/CUE/NRR - 2006/001, National F^ark Service. Washington, DC. 

THE BACKCROSS BREEDING PROGRAM OF THE AMERICAN CHESTNUT FOUNDATION 

Fred V. Hebard 
I American Chestnut Foundation Research Farms, 14005 Glenbrook Avenue, 

Meadowview, VA 24361-9703 USA (Fred@acf.org) 



Abstract: The blight resistance of oriental chestnut trees is being backcrossed into American chestnut 
using traditional plant breeding techniques. Progeny are screened for blight resistance by direct 
inoculation with the blight fungus, when they are old enough to survive inoculation, which is 3 or 4 years 
for trees with intermediate levels of blight resistance, and 1 or 2 years for trees with high levels of blight 
resistance. Trees are grown using intensive horticultural techniques. Probably the most unusual aspect of 
this breeding program in comparison to similar programs for crop plants is the large acreages over which 
trees are grown, and the fact that the objective is recovery of a genetically diverse species rather than an 
improved cultivar. Highly blight resistant progeny have been recovered from intercrosses of straight F|S, 
BiS and B2S, suggesting strongly that it should be possible to backcross blight resistance into American 
chestnut. Currently, two sources of blight resistance are being advanced to B3-F2. These are expected to 
begin producing progeny suitable for outplanting within 2 to 3 years. 



INTRODUCTION 

The American chestnut tree, Caslanea dentata (Marsh.) Borkh., has been destroyed as a dominant forest 
tree by a canker disease, chestnut blight, incited by Ciyphonectria parasitica (Murr.) Barr. The blight 
fungus was introduced into eastern North America around the turn of the 20" Century, probably in blight 
cankers on imported Japanese chestnut, C. crenata Sieb & Zucc, nursery stock (Metcalfe and Collins, 
1909). By 1950, the disease had killed almost all of the large American chestnut trees throughout their 
range. 

By 1930, when the American chestnut was thought to be doomed, attempts had begun to breed blight- 
resistant replacements. These attempts were abandoned, for the most part, around 1960, when no trees 
had been developed that combined the blight resistance of oriental chestnut trees with the large size of 
American chestnut trees (Jaynes, 1994). 

In 1961, what later proved to be viruses (Hillman e( al, 2000) were found infecting C. parastica (Grente, 
1961 ). The infected strains had been isolated from blight cankers on European chestnut trees, ('asianca 
saliva Mill., growing in Italy. The viruses reduced the virulence of the blight fungus enough that infected 
strains could no longer kill European chestnut trees. Additionally the viruses spread from one canker to 
another, resulting, apparently, in the protection of entire stands of European chestnut. When viruses were 
introduced into blight cankers on European chestnut in France, the disease there was ameliorated. This 
discovery led to efforts to control blight on American chestnut with these viruses, which continue today. 
To date, the results of this effort have not been entirely satisfactory (Anagnostakis, 1990). 

In 1981, Charles Burnham proposed that the blight resistance of oriental chestnut trees, primarily Chinese 
chestnut, Castama moUissima Blume, could be backcrossed into American chestnut. For American 
chestnut, this was a new method of plant breeding that had not been used in previous attempts to develop 
blight-resistant, timber-type chestnut trees. In 1983, The American Chestnut foundation was established 
as a not-for-profit corporation to help fund work on Burnham's proposal (Burnham et al., 1986). In 1989, 



61 



the foundation had accumulated sufficient resources to hire a part-time researcher at a new research farm 
in Meadowvievv, VA, in the heart of the range of the American chestnut tree. 

Subsequent to 1989, the foundation has grown to the point where it is supporting a large breeding effort in 
Meadowview, with four full-time workers tending trees on three farms totaling 130 acres. Additional 
workers are employed in Asheville, NC and at Penn State University' to assist volunteer breeding efforts 
at eleven state chapters. The administrative headquarters in Bennington, VT, also supports \olunteer 
breeding efforts in CT and VT. The purpose of this paper is to describe progress to date in this breeding 
program. 



MATERIALS AND METHODS 

Breeding Method 

To transfer blight resistance from Chinese to American chestnut, individuals of the two species are first 
crossed. The progeny from this cross, first hybrids, or FiS, usually are exactly one-half American and 
one-half Chinese chestnut. An F| is backcrossed to another American chestnut, decreasing the proportion 
of Chinese chestnut genes by a factor of one half, on average. The progeny of this second cross, the first 
backcross, are known as BiS. Two more backcrosses again decrease the proportion of Chinese chestnut 
genes by a factor of one half each time, to one-eighth followed by one-sixteenth, on average, with the 
remaining fraction of genes being from the American parent. 

At each step of backcrossing, resistant trees are selected by observing canker symptoms after inoculation 
of the progeny with the chestnut blight fungus (see below for details). The progeny also var\ in the 
fraction of Chinese genes remaining, and selection against Chinese morphological t\pe is made to 
accelerate recovery of the American type, using traits identified by Hebard (1995). Burnham estimated 
that three backcrosses to the American parent, with selection against Chinese morphological type, would 
be sufficient to recover trees that look and grow like the American chestnut of old. 

The F| trees, and any subsequent backcross progeny, would be heterozygous, at best, for the genes 
conferring blight resistance. Thus they would not be true breeding for blight resistance, throwing both 
susceptible as well as resistant progeny. To recover trees homoz\gous for blight resistance, third 
backcross trees are intercrossed among themselves, so the progeny have a chance of inheriting the genes 
for blight resistance from both parents. The progeny of this first intercross of third backcross trees are 
known as B3-F2S. 

Blight resistance is only partially dominant, so FiS and backcrosses are. at best, intermediate in resistance 
between the two parent species. High levels of blight resistance, comparable to those found in the 
Chinese parent, are only recovered after intercrossing F| hybrids and backcrosses. This facilitates 
recovery of trees reasonably homozygous for blight resistance, since they test out as more resistant than 
heterozygotes. 

To avoid inbreeding, and its consequent decrease in genetic diversit>, a different American chestnut 
parent is used at each step of backcrossing. Thus, in an ideal situation, four American parents are used to 
produce a third backcross tree. The third backcross progeny from a unique set of four American parents 
are termed a recurrent parent line or line for short. At the intercrossing stage, more than one line is 
needed in order to mininii/e sib crosses and their resulting inbreeding. Hebard (1993) estimated that 20 
lines would be needed to minimize loss of alleles from inbreeding. With four American parents per line, 
20 lines require 80 separate American parents. 



62 



In practice, only one line was used until the first backcross with the 'Graves' and 'Clapper' sources of 
blight resistance. These two first backcross trees then were crossed with 20 American parents to yield the 
second backcross generation, and with 20 additional parents to yield the third backcross. Thus the third 
backcross progeny are half first cousins rather than half third cousins. 

To ensure that the progeny from intercrossing third backcross trees are homozygous for blight resistance 
loci, only one Chinese chestnut parent is used to make a set of 20 lines. 

Sources of Blight Resistance 

The availability of the named first backcross, 'Clapper' (Little and Diller. 1964). and the undescribed 
'Graves' first backcross at the Connecticut Agricultural Research Station plantings in Hamden gave a 
jump start to the breeding program in 1989. These two first backcross trees were backcrossed again onto 
about 30 American chestnut trees each between 1989 and 1995 to yield second backcross trees, or B2S. 
Thirty American chestnut lines of third backcrosses were produced between 1996 and 2003 for both the 
'Clapper' and the 'Graves' lines. From 2001 until present, second generation third backcross progeny, or 
B3-F2S, have been collected and planted from intercrosses within sources of blight resistance. The 
Chinese chestnut grandparent of 'Graves' is an undescribed seedling known as 'Mahogany.' 

In 1989, breeding also was started with the Chinese chestnut cultivar, Nanking, crossing it with 20 
American chestnut trees to start 20 recurrent parent lines at Fi. Cultivar Nanking was chosen because it 
had shown the highest blight resistance of any Chinese chestnut tree evaluated by Headland and Griffin 
(1976) and was noted as having high blight resistance when first released. 

As available, other Chinese and Japanese chestnut trees, and Fi hybrids between these species and 
American chestnut, were crossed with American chestnut trees, in these later cases with only a few 
American chestnut trees rather than assembling 20 lines. Table ! lists the sources of blight resistance at 
their most advanced stage of backcrossing as of April, 2004, and the number of American parent lines at 
the most advanced stage. As indicated above, additional lines occur at less advanced stages of 
backcrossing for some sources of blight resistance. 



Table 1 . Oriental sources of blight resistance being used at The American Chestnut Foundation's 
Research Farms in Meadowview, VA, their most advanced stage of backcrossing into American chestnut 
and the number of American parent lines at that stage as of April, 2004. 



Source of Blight Resistance 


Stage of Backcrossing 


Number of American 
Parent Lines 


Clapper 


B3-F2 


12 


Mahogany 


B3-F2 


5 


Douglas 


B3 


2 


Nanking 


B3 


T 


Sleeping Giant South Lot Rl 1T14 


B3 


1 


Sleeping Giant South Lot R1T4 


B3 


1 


Sleeping Giant South Lot R1T7 


B3 


3 


Meiling 


B2 


1 


MusickChinese 


B2 


2 



63 



Greg Miller 72-211 


B| 


3 


mollissima? 


B, 




mollissimalO 


B, 




mollissima 13 


B, 




PI#104016 


Japanese Bi 




Dunstan seedling 


F, 




FP7284 


F, 




Greg Miller 65-18 


F, 


3 


Greg Miller 65-4 


F, 


6 


Kuling 


F, 


4 


Orrin 


F, 


4 


mollissima 1 1 


F, 


1 


mollissimalS 


F, 


1 


MAJ7Japanese 


Japanese F| 


1 


Jayne 


mollissima x pumila 


1 


AbbsVallev 


Chinese 




Altamont 


Chinese 




Armstrong 


Chinese 




Eaton 


Chinese 




MacBoyd 


Chinese 




MAJ 


Chinese 




MAJ4 


Chinese 




MAJ5 


Chinese 




Waynesboro 


Chinese 




mollissima 12 


Chinese 




mollissimal4 


Chinese 




mollissimal5 


Chinese 




mollissimal6 


Chinese 




mollissima 17 


Chinese 




mollissima 19 


Chinese 




mollissima20 


Chinese 




mollissimaS 


Chinese 




PI#7284 


Chinese 




PI#97853 


Chinese 




Richwood 


Chinese 




Wilkinson 


Chinese 




YardChinese 


Chinese 




FPGlenDaleID:GS 


Japanese 





64 



American Chestnut Parents 

In addition to the breeding at Meadow^/ iew, the American Chestnut Foundation also has an extensive 
network of state chapters staffed primarily by volunteers, and advised by staff officers stationed in North 
Carolina and Pennsylvania (Paul Sisco and Sara Fitzsimmons. respectively). The chapters have been 
crossing pollen of "Graves" and 'Clapper" second backcrosses from Meadowview onto local American 
chestnut trees to produce third backcross trees, for the most part. The intent is to produce a viable 
breeding population of 20 individuals for each source of blight resistance, adapted to the local conditions, 
and also to increase the genetic diversity of the breeding population, as originally proposed by Inman 
(1987). Table 2 depicts the number of third backcross trees in the various states as of 2004. 



Table 2. Number of third-backcross (B3) chestnut at TACF breeding orchards in 2004, with the number 
of sources of blight resistance and the number of American chestnut lines in the breeding stock. 



State 


Nuts or 
Trees 


Number of 
Sources of 
Resistance 


American 
Lines 


Maine 


1445 


2 


29 


Massachusetts 


3076 


2 


28 


Pennsylvania 


5350 


2 


36 


Maryland 


33 


1 


1 


Indiana 


1496 


1 


11 


Kentucky 


150 


2 





Virginia (Meadowview) 


5275 


8 


73 


North & South Carolina 


1049 


2 


9 


Tennessee 


745 


5 


6 


Alabama 


566 


1 


5 


Total 


19179 







Following Inman's recommendation (Inman, 1989), attempts have been made to limit the range of 
American chestnut parents to within 20 miles of each other in building local populations. This has been 
easier near Meadowview than elsewhere, since the required numbers of flowering chestnut trees can be 
found within such a small area. 

Pollination 

First hybrids and straight backcrosses are produced using the controlled pollination techniques described 
by Rutter ( 1991 ). Subsequent experience indicates that the best time to bag chestnut flowers for 
controlled pollination when the styles begin to emerge from the bur, rather than to assess the time by 
observing the onset of anthesis, as recommended by Rutter ( 1991 ). Experience also suggests that the slide 
technique using dried pollen described by Rutter ( 1991 ) to be more efficient than pollinating with fresh 
catkins. Flat surfaces other than microscope slides have been found preferable for applying pollen, such 
as the lid of the pollen container. In general, about one nut is produced per pollination bag placed over 
female flowers. 



65 



The intercross generations are produced by open pollination, where possible. Thus breeding orchards 
containing straight third backcross trees (B3) from one sources of blight resistance are isolated as much as 
possible from orchards with other sources of blight resistance or trees at other stages of breeding. 
Likewise, seed orchards, such as of B3-F2 trees, are isolated as much as possible from other orchards. A 
distance between orchards of about 1 kilometer is estimated to be sufficient to isolate orchards. Pollen 
from undesired trees also is eliminated by emasculation, pruning at ground level and removal of the , 
undesired trees. 

Cultivation 

The cultivation methods employed are standard orchard practices adapted to screening chestnut trees for 
blight resistance. Hebard ( 1991 ) discussed locating flowering American chestnut trees, and Hebard and 
Rutter ( 1991 ) outlined cultivation methods suitable for breeding orchards. Hebard (1994a) described the 
techniques for inoculating chestnut trees to test their blight resistance, and the orchard spacings used to 
grow trees. More recently, Hebard presented designs for seed orchards and methods for producing seed 
in them (2002) and methods for introducing additional sources of blight resistance into our chapter 
breeding programs (200 1 ). 

Orchards where backcross progenies are to be screened for blight resistance are arranged in completely 
randomized designs with controls consisting of 6 to 12 individuals each of pure American and pure 
Chinese chestnut trees, and their Fi hybrid. This experimental design was chosen because each genotype 
is unique, with no replication of genotypes. 

In a test of the response of trees of various ages to direct inoculation, the intermediate blight resistance of 
Fi hybrids as young as 1 year old was distinguished from the high resistance of pure Chinese and from 
susceptible pure American chestnut trees. However, F| hybrids did not sur\ ive the test unless the\ were 
at least 3 years of age. Thus straight second backcrosses. which also have blight resistance up to the 
intermediate level found in Fi hybrids, are screened for blight resistance when they are 3 or 4 years old. 
At those ages and under our growing condition, their diameter at breast height ( 1 .5 m) ranges from 3 to 
7.5 cm (1 to 3 inches) and their height from 3 to 5 m ( 10 to 15 feet). 

In order to avoid crowding prior to blight resistance screening, trees to be screened at 3 years of age are 
grown at a spacing of 1 .2 m (4 feet) within rows. Trees screened for blight resistance at 4 years of age are 
grown at a spacing of 2. i m (7 feet) within rows. Originally, straight backcross trees were screened for 
blight resistance at 4 years of age. Currently, straight backcross trees are screened for blight resistance 
when they are 3 years old. except for third backcross trees, which are screened when 4 years old (we did 
not wish to change methods for our most \aluable breeding material). Progenx of large. sur\ i\ ing 
American chestnut trees also are screened for blight resistance when the\ are 4 years old. To pro\ ide 
access for equipment, the between-row spacing in these orchards is 6 m (20 feet). 

Progenies expected to contain blight-resistant indi\ iduals. such as F^ generations, are screened for blight 
resistance when they are 1 or 2 years old. The blight-resistant progeny generally sur\ ive inoculation at 
that young age. These are spaced within rows at 30 or 60 cm (1 or 2 feet). The between-row spacing for 
Ft progeny varies from 2.1 to 6 m (7 to 20 feet) depending upon the location and intent of the test. 

Nuts are sown directly at orchard spacing. Prior to planting, orchard row s are subsoiled. plowed and 
rototillcd. and 31.75-|,im ( 1 .25-mil) black plastic mulch lain in 1.22-m-wide (4 feet) strips. Using handled 
bulb planters, holes are drilled through the mulch into the soil and filled with a mix of one-third each 
ground, milled peat moss, perlite and coarse vermiculile. Nuts are planted I -cm deep (0.5 inches) and 
protected from voles with aluminum cylinders 25.4-cm tall (10 inches) and 5 to 7 cm wide (2 to 3 inches). 
After planting, the cylinders are jammed down around the nuts to a depth of about 5 cm (2 inches). The 



66 



aluminum is painted to reduce aluminum toxicity should it dissolve into the soil. Soil is mounded around 
the cylinders to prevent them from being blown away by wind. Styrofoam cups are inverted over 
cylinders until shoots emerge from the cylinders. At that point, the bottom of the cup is removed, and the 
cup replaced, to diminish breaking of the young shoots on the edge of the cylinders. 

The seedlings generally outgrow the width of the cylinders during their third growing season. At the 
beginning of the third growing season, the cylinders are removed. The mulch also is removed to reduce 
vole damage. Prior to this time, the cylinders prevent vole damage. Voles can be harbored under mulch. 

While black plastic mulch is in place, trees are fertilized with soluble fertilizer with a major nutrient 
composition of 30-10-10 (N-P-K) plus cationic trace elements (MirAcid^'^ or equivalent). Liquid 
fertilizer is used in order to place the fertilizer under the impermeable mulch. Approximately 2 liters (2 
quarts) of fertilizer solution is applied every 2 weeks between mid May and early August. The fertilizer 
concentration is 3.26 ml per liter ( 1 .25 tablespoons per gallon of water). Fertilizer is pumped directly 
down the cylinders or applied through a drip irrigation system. Once plastic mulch is removed, granular 
fertilizer is broadcast around the trees. The rate for granular fertilizer usually is 224 kg per hectare (200 
lbs per acre) of N as ammonium nitrate and diammonium phosphate, 67 kg per hectare (60 lbs per acre) of 
P as diammonium phosphate and 67 kg per hectare of K as potash. These amounts are applied twice a 
year, in mid May and late June. In seed orchards, to avoid having to apply liquid fertilizer underneath 
plastic mulch, landscape fabric is used for mulching and granular fertilizer is broadcast at the above rates. 
The rates were formulated from soil and foliar mineral analysis for the soils typical of Meadowview and 
might differ on other soils. The rates also are adjusted depending upon the results of soil mineral 
analysis. 

On trees 5 years of age and younger, weeds are managed with herbicides and mulch. In general, no weed 
management is performed on trees older than 5 years of age, other than mowing. Currently, in P pril, 
Surflan''^'^ A.S. (oryzlin) is applied at 9.35 liters per hectare (4 quarts per acre), simizine 4L at 7.02 liters 
per hectare (3 quarts per acre) and Roundup Ultra^'^ (glyphosate) at 3.07 liters per hectare (42 oz per 
acre). A supplemental spray of Roundup Ultra^'^ at 3.07 liters per hectare (42 oz per acre) is applied in 
July to trees younger than 3 years old. These herbicides are applied as a directed spray using TeeJet^"^ 
8005 standard flat-fan nozzles operated at 2.07 bars (30 psi) in a water solution of 608 liters per hectare 
(65 gallons per acre). The combination of low pressure with high volume spray nozzles increases droplet 
size, reducing drift. A strip 152.4 cm wide (3 nozzles at 50.8-cm or 20-inch spacing, 45.72 cm or 18 
inches above the ground) is sprayed down each side of a row. The nozzle closest to the trees is directed 
with a hand wand, the other two nozzles are mounted on the boom of the spray rig. 

Grass strips are maintained between rows to reduce erosion. Fire hazard is reduced by regular mowing 
with rotary cutters. In B3-F2 seedling seed orchards, which are sown at much higher densities (0.3 \ 2.1 
m, 1 X 7 feet), maintenance is performed with a riding lawn mower. Weeding of seedling seed orchards is 
done as above, but using a 25-gallon tow-behind sprayer attached to the lawn mower rather than a 65- 
gallon herbicide spray rig mounted on the three-point hitch of the standard orchard tractors used in the 
larger orchards. Only two nozzles are used in seedling seed orchards. The lawn mower-mounted nozzle 
is attached to the front of the mower. The mower operator also can manipulate a hand wand fairly easily 
on the lawn mower, whereas on the larger orchard tractors it is best if the hand nozzle is operated by a 
person walking behind. A pressure regulator needs to be added to most tow-behind sprayers. Their 
pumps are driven by electric motors powered from the lawn mower's electrical system, whereas the 
power take off drives the pumps on the orchard tractors. Thus it is important that the lawn mower 
produce enough electric current to power the pump. 

Using an airblast sprayer, aphids are controlled with a single application of dormant oil during bud break 
at 56 liters per hectare (6 gallons per acre) in 2807 liters per hectare (300 gallons per acre) of water 



67 



solution. In July, Japanese beetles are controlled with 2 to 3 applications of Sevin XLR Plus''"" at 5.8 
liters per hectare (0.625 gallons per acre) in 935 liters per acre (100 gallons per acre) of water solution. 
Spray amounts have been reduced considerably by employing a Durand-Wayland Smart Spray 1000^'^ 
attached to a Durand-Wayland model AF500CPS airblast sprayer. This device cuts otT banks of nozzles 
depending upon tree height and occurrence. 

The pesticide application methods, composition, and rates were formulated in consultation w ith extension 
specialists from the Virginia Polytechnic Institute and State University and the "Sprav Bulletin for 
Commercial Fruit Growers," which is issued annually (Virginia, West Virginia & Mar\ land Cooperative 
Extension Services, 2004). 

Straight backcross trees have been irrigated in the year of inoculation during dr>' years. Since the vear 
2000, all young chestnut trees have been irrigated, except B3-F2 seedlings, using a drip irrigation system. 
Soil moisture is maintained at field capacitv (about 10-20 kiloPascals of soil moisture deficit). We plqn 
to not irrigate B3-F2 seedling seed orchards. 

Trees are not pruned for shaping or for removal of lower branches, as is often done in commercial fruit 
and nut orchards to facilitate passage down the rows and weeding with herbicides, among other 
objectives. Not pruning results in a crown that extends to the ground on the trees (and necessitates a 
second person walking behind the herbicide sprayer to prune off portions of branches that are spraved 
inadvertently). This larger crown may promote early and heavier bearing. For the most part, our trees 
produce male catkins when they are 2 to 4 years old and bisexual catkins when they are 3 to 5 years old. 
This early flowering also has been seen in other hardwood trees grown under intense cultivation (Wright. 
1976). 

Using the above methods, the trees at Meadowview have averaged 0.56 m tall after one growing season, 
1 .5 m (5 feet) tall after two, 2.4 m (8 feet) after three, and 3.7 m ( 1 2 feet) after four grow ing seasons. 
There can be considerable variation in height growth within orchards and between grow ing season, 
genotype and location. 

Screeninu for Blight Resistance 

The cork-borer, agar-disk method is used to inoculate chestnut trees with the blight fungus (Griffin, el al. 
1983). Agar disks are obtained from the margins of growing cultures that have not reached the edge of 
the Petri plate. Inoculations are performed in early June. This is the earliest in the season when cool 
weather (daily high temperatures below 15 to 20 C) can be avoided reliably. Cool weather occurs every 
few years in late May in Meadowview and can lead to inoculation failure. 

Two strains of the blight fungus are used, known as Epl55 and SGI 2-3. Epl55 is a widely used strain of 
the blight fungus (ATCC 38755), while SGI 2-3 was isolated near Meadowview by the author. When 
tested for pathogenicit}' in American chestnut, the distribution of lengths of cankers incited by \ irulent 
strains of the blight fungus follows a bell-shaped curve; it is approximatclx normall\ distributed, and 
variances are equal for the various canker lengths (Gritfm, et al. 1983). When replicated fwc times each 
over 3 years, or 15 total replicates. Ep 155 was among the most pathogenic of 21 tested virulent strains, 
having significantly (p < 0.05) larger cankers than six of the least pathogenic test strains. Likewise. SGI 
2-3 was among the least pathogenic of the 21 tested strains, having significantly smaller cankers than 
seven of the most pathogenic test strains. 

Blight resistance can be determined quantitatively by measuring the length and width of cankers. Canker 
depth or superficiality is not detenu incd at Meadowview since the intermediate to ver\ high levels of 
blight resistance being sought can be distinguished using length and width measurements alone. Until 



68 



1999, the length and width of cankers was measured on all tested trees. Because this was taking too much 
time, beginning in 1999. blight resistance in most tests was determined using a qualitative assessment. 

The qualitative assessment is based on the following observations. In general, 1 year after inoculation, 
SGI 2-3 incites small cankers (2-3 cm long) on trees with intermediate levels of blight resistance or 
higher. It incites medium-sized cankers (3-6 cm long) on trees with low levels of blight resistance, and 
large cankers (> 6 cm long) on normal American chestnut trees. In contrast, Ep 155 incites large cankers 
on trees with intermediate levels of blight resistance or less, medium-sized cankers on trees with high 
levels of blight resistance, and small cankers on trees with very high levels of blight resistance. Thus five 
blight resistance classes can be distinguished on trees inoculated with both strains. This is depicted 
visually in Table 3. 



Table 3. Blight resistance classes distinguished qualitatively by various canker length classes for two 
strains of Ciyphonectria parasitica one year after inoculation in early June. 



Numeric blight 
resistance class 


Verbal blight 
resistance class 


Length (cm) of canker 
incited by 


Ep 155 


SGI 2-3 


1 


highly blight resistant 


2-3 


2-3 


2 


blight resistant 


3-6 


2-3 


3 


intenned lately blight resistant 


>6 


2-3 


4 


slightly blight resistant 


»6 


3-6 


5 


not blight resistant or 
susceptible 


>» 6 


>6 



Table 3 depicts idealized canker lengths for various blight resistance classes seen in average years. 
Depending upon the season, slightly blight-resistant trees might show small SGI 2-3 cankers or blight- 
resistant trees might show large Ep 1 55 cankers. Additionally, the responses to the two strains do not 
always move in parallel with each other. These various unusual patterns of response can be detected by 
the response of the pure American and Chinese chestnut trees and their Fi hybrids planted as control trees 
in the orchard and the scale adjusted accordingly. 

In addition to artificial inoculation, trees in Meadowview also are exposed to naturally occurring 
inoculum. Blight incidence due to natural infections on straight backcross progeny exceeds 50% by the 
beginning of the fifth growing season, when trees are four years old. When screening artificially 
inoculated trees for blight resistance, the severity of these naturally occurring cankers is considered in the 
overall assessment of a tree. Thus, while only two strains of the blight fungus are used for direct 
inoculation, a larger number of strains is involved in the overall assessment. 



RESULTS AND DISCUSSION 

Recovery of highly blight-resistant backcross progeny at F^ 

The first screening of progeny segregating for blight resistance in Meadowview occurred in 1993. One 
set of progeny consisted of B1-F2S obtained from reciprocal crosses of the 'Graves' and 'Clapper' trees. 
A second set of progeny consisted of straight F2S obtained from a one-way cross of two Fis. The Fi 



69 



parents were half sibs from crosses of the 'Mahogany' Chinese chestnut tree with pollen from two 
American chestnut trees. A third set of progeny segregating for blight resistance consisted of straight B2S 
composited from three crosses of pollen from the "Graves" tree onto three American chestnut trees. The 
trees were 2 years old when inoculated in June, 1993, and the data in Table 4 summarize canker 
dimensions when measured in September, 1993. Each tree was inoculated once with strain Ep 155 and 
once with strain SGI 2-3, using the cork borer, agar-disk method with holes 2 mm in diameter. Highly 
blight-resistant progeny were recovered from the F2 and the B1-F2 crosses, and progeny with intermediate 
levels of blight resistance were recovered from the B: crosses. The B|-F: crosses may have had higher 
blight resistance than the straight F^s. Figure 1 depicts one of these highly blight-resistant B1-F2S. 



Table 4. Mean and standard deviation and distribution of canker size classes (mean length and width of 
cankers incited by two strains of the blight fungus) for straight F2, B1-F2 and B2 American x Chinese 
chestnut progeny and controls. 



Cross Type 






Canker 


size class (cm) 






Mean 


Standard 
deviation 


1.0 to 
2.6 


2.6 to 

4.2 


4.2 to 
5.8 


5.8 to 

7.4 


7.4 to 
9.0 


9.0 to 
10.6 


10.6 
to... 


Seedling American 










3 


5 


2 


9.6 


1.1 


Fi 'Nanking' 








2 


4 


3 




8.4 


1.0 


Seedling Chinese 




2 


7 


3 








5.2 


1.0 


"Meiling" Chinese 




1 


2 


2 








5.5 


1.1 


'Nanking' Chinese 


3 




2 










2.9 


1.4 


F2 'Mahogany' 




5 


23 


48 


48 


29 


15 


7.7 


1.9 


B1-F2 "Clapper' x 'Graves' 


4 


25 


84 


116 


112 


54 


4 


6.9 


1.9 


B2 "Graves' 






2 


4 


15 


26 


6 


9.1 


1.5 



Three-year-old B2-F2 progenies from controlled crosses between selected straight B2S (backcrossed to 
American chestnut) were inoculated in June, 2003, and cankers measured in November. "Clapper' B2-F2 
progeny were from a single cross between two half sibs, while "Graves" B2-F2 progeny were a composite 
of three crosses between half sibs. Depending upon their size, these trees were inoculated once or twice 
each with strains Ep 155 and SGI 2-3, using the cork borer, agar-disk method, but the holes were 4 mm in 
diameter. A larger cork borer and number of inoculations were used in 2003 than in 1993 because 2003's 
3-year-old trees were larger than 1993's 2-year-old trees. Again, highlv blight-resistant progeny were 
recovered, this time from second backcross F2S (Table 5). Thus, not only could highlv blight-resistant 
progeny be recovered by intercrossing Fi interspecific hybrids or by intercrossing first or second 
backcrosses to American chestnut, but high levels of blight resistance were retained through the second 
backcross. These results suggest very strongly that the blight resistance of Chinese chestnut can be 
backcrossed into American chestnut. 

Canker sizes were smaller in the 2003 than in the 1993 test, possibly because of cooler, wetter weather in 
the later year, so there was not as much separation of canker sizes among the controls. However, the 
cankers on some of the B3-F2 progeny have remained small through the 2004 growing season, as 
illustrated in Figure 2. An earlier test, performed in 1999 on open-pollinated progeny of "Clapper" B2S, 



70 




Figure 1. Highly blight-resistant Chinese to American B1-F2, 13 years old, 
with Cryphonectria parasitica. The tree is to the left of and behind the dog. 



years after inoculation 



presumably pollinated by other 'Clapper' B2S, gave results similar to those presented in Table 5 (Hebard 
ef ah 2000). 

Blight resistance in straight backcrosses 

Tables 6, 7, and 8 report typical results of rating straight second and third backcross trees for blight 
resistance. An entire family derived from a second backcross tree has not yet been rejected based on the 
performance of its third backcross progeny. In general, the blight resistance of third backcross progeny is 
comparable to that observed in second backcross trees, again supporting the inference that there is no 
diminution of resistance as backcrossing proceeds. 

Family effects have occurred in second backcross progeny fathered by both the 'Graves' and 'Clapper' 
trees, where the American mother of second backcross progeny influenced their phenotypic blight 
resistance. This is illustrated in Table 9, where the Bu3ClC x 'Clapper" family had cankers closer in size 
to cankers on Chinese chestnut than on FiS or Americans. It is unclear whether or not the Bu3C IC 
American parent was contributing genes for blight resistance by itself or contributing genes that 



71 



Table 5. Distribution of canker size classes (mean length and width of cankers incited by two strains of 
the blight fungus) for B2-F2 American x Chinese chestnut pre^geny and controls. ~ 



Cross Type 


Canker size class (cm) 


Mean 


Standard 
deviation 


1.0 to 
2.0 


2.0 to 
3.0 


3.0 to 
4.0 


4.0 to 
5.0 


5.0 to 
6.0 


6.0 to 

7.0 


7.0 to 
8.0 


Seedling American 






4 


2 


2 


2 


1 


5.0 


1.4 


F) 'Nanking' 




1 


2 


3 


1 






4.1 


1.0 


Seedling Chinese 


3 


3 


3 


6 








3.3 


1.2 


B2-F2 'Clapper' 


3 


11 


15 


37 


16 


12 


3 


4.5 


1.4 


B2-F2 'Graves' 


3 


11 


21 


31 


14 


14 


1 


4.4 


1.3 



Table 6. Blight resistance ratings of 'Clapper' and "Graves* second backcross trees and controls in 1999. 



Cross type 


Blight resistance rating 


1 


2 


3 


4 


5 


Seedling American 








2 


3 


Fi 'Nanking' 




4 








Seedling Chinese 


3 


5 








'Nanking' Chinese 


1 


1 








B2 'Clapper' 




5 


27 


29 


12 


B2 'Graves' 




3 


42 


47 


25 



Table 7. Blight resistance ratings of 'Clapper' third backcross trees and controls in 2000. 



Cross type 


Blight resistance rating 


1 


2 


3 


4 


5 


Seedling American 








3 


3 


Fi 'Nanking' 




2 


10 






Seedling Chinese 


3 


2 


1 






B3 'Clapper" 


1 


19 


139 


383 


95 



Table 8. Blight resistance ratings of 'Graves" third backcross trees and controls in 2001 



Cross type 


Blight resistance rating 


1 


2 


3 


4 


5 


Seedling American 






1 


3 


8 


F) 'Nanking' 




2 


5 






Seedling Chinese 


7 


8 








B3 'Graves' 






124 


124 


122 



72 



modulated the expression of blight resistance genes from Chinese chestnut. The Bu3ClC tree itself did 
not appear to have more blight resistance than t>pical American chestnut trees; it died from blight the 
year after this cross was made, like most of the other American chestnut trees at that site. 




Figure 2. Left, chestnut blight cankers after two growing seasons on a highly blight-resistant 
'Clapper' B2-F2. Top left, canker incited by strain SGI 2-3. Bottom left, canker incited by strain Ep 
155. Right, 4-year-old 'Clapper" B2-F2. Similar cankers on blight-susceptible American chestnut 
would be expected to exceed 40 cm in length; these cankers were 2 to 3 cm long. 



73 



Table 9. Distribution of canker size classes (mean length and width of cankers incited by two strains of 
the blight fungus) for progeny of second backcrosses of the "Graves' and "Clapper" first backcross trees to 
American chestnut and controls, in 1998. 



Cross type 


Canker size class (cm) 


2 to 4 


4 to 6 


6 to 8 


8 to 10 


lOto 12 


12to 14 


14 to 16 


Seedling American 










4 


1 




F] "Nanking' 






1 


3 


1 






Seedling Chinese 




3 


4 










"Nanking' Chinese 





2 












Bu2B2C X "Clapper" 






2 




3 


2 




Bu2B3C X 'Clapper' 






3 


4 


4 


1 




Bu3ClCx "Clapper" 




15 


33 


8 








BulClG X 'Graves' 






4 


8 


17 


10 


1 


BulC2G X 'Graves' 






2 


1 


1 


1 




Bu3BlGx 'Graves' 








1 








Bu3B2G X 'Graves' 

















Bu3C3C X 'Graves" 




4 


8 


15 


25 


19 


2 


Bu3DlGx 'Graves' 






1 




2 






Bu3FlG X 'Graves' 










1 


1 




Bu3F5G X 'Graves' 






2 


2 


5 


2 




Bu3RlGx 'Graves' 






-) 


7 


13 


4 


1 



Number of genes conditioning blight resistance 

The standard deviations of canker size in Table 4 were greater for the progeny expected to be segregating 
for blight resistance than for the controls, and, for the F2S, were compatible \\ ith models for one or two 
incompletely dominant genes controlling blight resistance, using Wright's method for estimating the 
number of factors controlling a segregating trait (Falconer, 1 960, p 2 1 8). (In this computation, the total 
genetic variance of the F^s was substituted for the additive genetic variance; the fomier was computed h\ 
subtracting the mean variance of the controls from the variance of the F:S. The broad sense heritabilit\ 
calculated from these variances was about 70%). The distributions of canker size in segregating progeny 
in Table 4 were compatible with the distributions of canker size expected for two or three incompletely 
dominant genes of equal effect on blight resistance, among other models for gene action. Similar models 
with more than three factors or fewer than two did not t1t the observed values (chi-square p«0.05). The 
expected distributions were constructed from the mean response for the control trees, assuming a normal 
distribution of canker size with the average standard deviation of the controls shown in fable 4; missing 
cells, such as for trees with only one allele for resistance, were estimated by linear interpolation between 
the relevant observed values. Unfortunately, vegetatively propagated (grafted) individuals of "Mahogan\ " 
were not available for inclusion in the test, nor the actual Fi parents: otherwise stronger inferences might 
have been possible concerning the mode of inheritance of blight resistance. Subsequent experience 
suggests that "Mahogany" has a high le\el of blight resistance, comparable to that of "Nanking." This 



74 



suggests in turn that two genes are involved in blight resistance. A three-gene model would be more 
compatible with the data if Mahogany' Chinese chestnut had a "normal" level of blight resistance like 
the 'Meiling' and seedling Chinese in Table 4 rather than the high level of blight resistance observed in 
'Nanking. "Kubisiak. et a/. (1997) prepared a genetic map of the 'Mahogany" F^s whose canker sizes are 
shown in Table 4. Their results indicated that three regions of the genome were associated with blight 
resistance. The Kubisiak, ef al. ( 1997) map was constructed with randomly amplit'ied polymorphic 
deoxyribonucleic acid markers (RAPDs), restriction fragment length polymorphic markers (RFLPs), and 
isozymes. Subsequent genotyping of the mapping population with markers based on simple sequence 
repeats (SSRs) indicated that 17 of the 185 progeny were outcrosses, not pollinated by the supposed male 
parent (Sisco. Kubisiak and Hebard, unpublished). These individuals are not included in Table 4. One of 
the three regions of the genome previously associated with blight resistance (located on Kubisiak el al. 
( I997)'s linkage group G) was no longer associated with blight resistance in the revised mapping 
population. Molecular mapping of backcrosses of 'Mahogany" Fis to American chestnut also suggested 
that the same two regions of the genome condition blight resistance (Kubisiak and Hebard, unpublished). 
The molecular mapping data thus supported a model of two incompletely dominant genes conditioning 
blight resistance in these progeny. 

Highly blight resistant 'Clapper' x 'Graves' B1-F2 individuals were test crossed to American chestnut to 
determine whether or not they were homozygous for blight resistance. Screening of these 'Clapper' x 
'Graves' test cross progenies indicated that they were segregating for blight resistance (data not shown). 
and hence that the Bj-F: parents were not homozygous. This finding suggests that some of the genes 
conditioning blight resistance in 'Clapper' and 'Graves' are at different loci. Highly blight-resistant 
'Mahogany' F2 progeny also had been test crossed to American. Unfortunately, all of the test-crossed 
individuals turned out to be outcrosses, as indicated by the SSR markers, invalidating this second set of 
tests. 

There are numerous patterns of inheritance possible when a trait is controlled by more than one gene, 
including complementary inheritance, epistasis, etc (Grant, 1975). The model here of two incompletely 
dominant genes of equal effect is only one among these models, albeit one that fits the data. If further 
improvement of backcross chestnut trees for blight resistance is necessary beyond the B3-F2 stage of 
breeding, it might be best to use breeding methods for quantitative traits, such as recurrent selection. 

The fact that the variance or range of canker sizes for the Fj controls in Tables 4 to 9 were similar to those 
of the pure species indicates that 'Nanking* Chinese chestnut t.^ees are homozygous for blight resistance. 
Similar data suggest that the named Chinese chestnut cultivars Orrin and Meiling. and the unnamed 
cultivars of Greg Miller, 64-4 and 72-21 1, likewise are homozygous for blight resistance. 

Outbreeding and inbreeding depression 

Not infrequently, specific Chinese x American chestnut crosses fail to produce nuts. Sometimes, nuts are 
produced, but fail to grow after germinating a radicle. These failures may be considered extreme 
instances of outbreeding depression. Chinese x American F| hybrids that do germinate often exceed pure 
species in size up to 1 to 20 years after planting, exhibiting hybrid vigor. For instance, after three 
seasons of growth. Fi hybrids in four orchards were significantly (p < 0.0001 ) taller than pure species, 
having a least squares mean height of 2.2 m versus 1 .8 m for the pure species. The Fi hybrids also were 
significantly taller than any of the individual pure species. 

The 'Mahogany' F2S of Table 4 came from the only intercross of Chinese x American F| hybrids that has 
yielded well (greater than 1 .0 nuts per pollination bag). Other F; intercrosses have yielded fewer than 0.6 
nuts per pollination bag. sometimes much less. Attempts to use Chinese x American Fi hybrids to 
pollinate American or Chinese chestnut trees also have produced low yields, in general. Even some 



75 



intercrosses among half-sib B^s have yielded sound nuts that failed to produce seedlings. The failures of 
some of these more advanced crosses may be due to inbreeding depression rather than outbreeding 
depression. The failures (and pollen contamination in our early crosses) bedeviled attempts to repeat the 
early experiments. Similar failures also may have bedeviled attempts of earlier researchers to test 
hypotheses regarding the inheritance of chestnut blight resistance. 

As mentioned previously in the section on blight resistance, the "Clapper x "Graves' Bi-F^s of Table 4 
had more apparent blight resistance than the Mahogany F2S. They also grew to be larger, more \ igorous 
trees, perhaps because they did not suffer from inbreeding depression and/or had hybrid \ igor (four- 
hundred, nineteen "Clapper x Graves" and 'Graves x Clapper" progeny had a mean height at the end of the 
1993 growing season of 2.43 m while 191 "Mahogany" F2s had a mean height of 2.13 m, significantly 
shorter, p<0.0001 ; a similar trend, p=0.001, was observed in 1992. prior to inoculation). The relative 
contributions of general vigor versus specific genes for blight resistance to the greater phenotxpic blight 
resistance of the "Clapper" x "Graves" Bj-F^s are unclear. 

Summary 

In sum, we have been able to recover highly blight-resistant chestnut trees after backcrossing blight 
resistance from Chinese into American chestnut for two cycles of backcrossing. Three cycles of 
backcrossing are expected to produce chestnut trees that, for the most part, look and grow like American 
chestnut. We currently are starting to test the blight resistance of second-generation, third backcross trees 
(B3-F2s), and currently expect some of them to have high levels of blight resistance. By 2008. we hope 
to begin planting their progeny (B3=F3s) back into the forest to confirm these expectations and to begin 
restoring the American chestnut tree to Appalachian forests. 



LITERATURE CITED 

Anagnostakis, S. 1990. Improved chestnut tree condition maintained in two Connecticut plots after 
treatment with hypovirulent strains of the chestnut blight fungus. For. Sci. 36: 1 13-124. 

Burnham, C.R., P. A. Rutter, and D.W. French. 1986. Breeding blight-resistant chestnuts. Plant Breed. 
Rev. 4:347-397. 

Falconer, D.S. 1960. Introduction to quantitative genetics. Ronald, New York. 365 p. 

Grant, V. 1975. Genetics of Flowering Plants. Columbia Universit>' Press. New York. 514 p. 

Grente. J. 1961 . Observations sur le comportement des plants de chataignier apres inoculation de 
V Endolhia parasitica Ann. Epiphyties 12:65-70. 

Griffin. G.J.. F.V. llcbard. R.W. Wendt. and J.R. Elkins. 1983. Survival of American chestnut trees: 
evaluation of blight resistance and virulence in Emiothia parasitica. Ph>topatholog\ 73:1084-1092. 

Headland, J.K.. G.J. Griffin, R.J. Stipes, and J.R. Elkins. 1976. Severity of natural Emiothia parasitica 
infection on Chinese chestnut. Plant Dis. Rep. 60:426-429. 

licbard. F.V. 1991 . Locating Howering American chestnut trees. J. Am. Chestnut Found. 5:98-100. 

Hebard, F.V. 1994a. The American Chestnut Foundation breeding plan: beginning and intermediate steps. 
J. Am. Chestnut Found. 8:21-28. 



76 



Hebard, F.V. 1994b. Inheritance of juvenile leaf and stem morphological traits in crosses of Chinese and 
American chestnut. J. Hered. 85: 440-446. 

Hebard, F.V. 2001. Meadowview Notes 2000-2001. J. Am. Chestnut Found. 15:7-17. 

Hebard, F.V. 2002. Meadowview Notes 2001-2002. J. Am. Chestnut Found. 16:7-18. 

Hebard, F.V., and P. A. Rutter. 1991. Growing chestnut trees from seed. J. Am. Chestnut Found. 5:1 10- 
113. 

Hillman, B.I., D.W. Fulbright, D.L. Nuss, and N.K. Van Alfen. 2000. Hypoviridae. P. 5 1 5-520 in Virus 
Taxonomy: Seventh Report of the International Committee for the Taxonomy of Viruses, van 
Regenmortel, M.H.V., et al. (eds.). Academic Press. New York. 

Inman, L.I. 1987. Proposed strategies to preserve and restore the American chestnut. J. Am. Chestnut 
Found. 2:6-9. 

Inman, L.I. 1989. Simultaneous breeding of the American chestnut for many traits. J. Am. Chestnut 
Found. 4:16-17. 

Jaynes, R.A. 1994. Reflections. P. 45-46 in Proceedings of the International Chestnut Conference, 
Double, M.L. and W.L. MacDonald (eds). West Virginia University Press, Morgantown. 

Kubisiak, T.L., F.V. Hebard, CD. Nelson, J. Zhang, R. Bernatzky, H. Huang, S.L. Anagnostakis, and 
R.L. Doudrick. 1997. Molecular mapping of resistance to blight in an interspecific cross in the genus 
Castanea. Phytopathology 87:751-759. 

Little, E.L., and J.D. Diller. 1964. Clapper chestnut, a hybrid forest tree. J. For. 62:109-1 10. 

Metcalf, H., and J.F. Collins. 1909. Present status of the chestnut bark disease. USDA Bull. 141, Part 5, p. 
45-54. 

Rutter, P. A. 1991. Quick guide to making controlled pollinations of chestnut. J. Am. Chestnut Found. 
5:93-97. 

Virginia, West Virginia and Maryland Cooperative Extension Services. 2004. 2004 Spray bulletin for 
commercial fruit growers, publication 456-419. Virginia Polytechnic Institute and State University, 
Blacksburg, VA. 148 p. 

Wright. J.W. 1976. Introduction to forest genetics. Academic Press, New York. 463 p 



77 



78 



Steiner, K. C. and Carlson, J. E, eds. 2006. Restoration of American Chestnut To Forest Lands - Proceedings of a 
Conference and Workshop. May 4-6, 2004, The North Carohna Arboretum. Natural Resources Report 
NPS/NCR/CUE/NRR - 2006/001, National Park Service. Washington, DC. 



BLIGHT RESISTANCE TECHNOLOGY: TRANSGENIC APPROACHES 

William A. Powell', Scott A. Merkle", Haiying Liang"\ and Charles A. Maynard' 
College of Environmental Science and Forestry, 
SUNY, Syracuse, NY 13210 USA (wapowell@esf.edu) 
Daniel B. Wamell School of Forest Resources, University of Georgia, Athens, GA 30602 USA 
^Pennsylvania State University, University Park, PA 16802 USA 



Abstract: The technology needed to produce a blight-resistant, transgenic American chestnut tree has 
come to fruition. Many technical hurtles have been overcome and the first transgenic American chestnut 
trees are expected to be in the greenhouse within one to two years. These first trees will contain an 
oxalate oxidase transgene from wheat, under control of a regulated promoter from soybean. This will be 
the first of several putative resistance-enhancing transgene constructs to be tested for its ability to confer 
chestnut blight resistance. During the field trial phase of this research, we hope to include a few 
educational research plots accessible to the public under controlled conditions. In this way we can 
enhance the transparency of this project by allowing the general public to see first hand the results of this 
research. Once the transgenic trees are approved for release, they will enter a restoration program to 
rescue as much of the remaining genetic diversity in the surviving American chestnuts. This research 
should also help pave the way for the restoration of other threatened tree species through the use of 
biotechnology. 



INTRODUCTION 

Due to the movement of plant materials around the world, many exotic diseases and pests have threatened 
North American forests over the past century. Tree diseases such as white pine blister rust, chestnut 
blight, beech scale complex, Dutch-elm disease, butternut canker, dogwood anthracnose, just to name a 
few examples, have caused significant losses of trees in our forests and in urban settings. Although the 
United States has established several regulatory safeguards, diseases, such as the newly discovered 
sudden oak death, sometimes get through. It is likely that new introductions of diseases and pests will 
continue into the future. In an effort to restore species from past and ongoing epidemics, researchers are 
beginning to apply biotechnological techniques to forest species. In our labs, we are taking a transgenic 
approach to enhance pathogen resistance in American elm and in American chestnut. This report will 
describe the progress made to date and discuss the possible future of blight-resistant, transgenic American 
chestnut trees. 



TRANSGENE DESIGN 

When choosing transgenes to enhance pathogen resistance, and to play a vital role in the restoration of a 
valuable tree species such as the American chestnut {Castanea dentata (Marsh.) Borkh.). several 
considerations need to be taken into account. First, if the tree produces an eatable product, like the nut of 
the American chestnut, it must be equally as safe to eat for humans and wildlife as the nuts from non- 
transgenic trees. Second, the transgenic tree should have effective and durable resistance to the blight. 
Third, the transgenic tree should retain all the positive traits of the American chestnut so that it can be 
reestablished in its natural niche in the environment. Fourth, the transgenic tree should be amenable to a 
restoration program. For example, it should be able to recapture a significant portion of the remaining 
genetic diversity in the surviving population. Lastly, it needs to be acceptable to the general public, i.e. 



79 



the majority of the public should view the trees positively. In our transgene designs, we have considered 
all these aspects when choosing what to use. ~ ^ 

Several genetic components needed to construct a variety' of putative resistance enhancing genes are" 
currently available (Powell et al. 1995; Powell and Maynard 1997: Powell et al. 2000; Liang et al. 2001; 
Connors et al. 2002; Connors et al. 2002; Liang et al. 2002). More putative resistance enhancing genes 
are being reported each year and someday the resistance genes from the Asian chestnuts trees might be , 
cloned, enhanced, and used to transform American chestnut trees. But to save time and space, this report 
will focus only on the transgene construct that will be used first to produce a transgenic American 
chestnut. If this construct fulfills all the necessar> criteria for producing a blight-resistant American 
chestnut, then this might be the only construct needed, but if this transgenic construct is not as effective as 
needed, many other genes and gene promoters are available. The first transgenic American chestnuts will 
contain a three-gene cassette as show in the binary vector pVSPB-OxO (fig. 1 ). 

This construct contains an oxalate oxidase (OxO) 
encoding gene from wheat (Lane et al. 1986; Lane 
1994). This gene was selected because it comes from a 
familiar plant that is consumed by most Americans ever> 
day and therefore brings with it a sense of public 
acceptability. It is also being researched for use to 
enhance pathogen resistance in other transgenic crop 
species, which should help with the regulatory approval 
as government reviewers become familiar with this 
transgene" s properties. Lastly, its mechanism for 
enhancing resistance appears to be well suited for 
producing effective and durable resistance to the 
chestnut blight. 



VSPB Promo t, 




Uki-3-Pr«a>otrr 



BamHI 
35S-nngfp5-er 



Pad 



Figure 1 . Plasm id map of the binary 
vector pVSPB-OxO. 



Oxalate oxidase catalyzes the degradation of oxalic acid 
into H2O2 and COo. Similar enzymes have been found in 
several plant species such as barley and wheat and are 
expressed during germination, stress, and fungal 
infection (Dumas et al. 1995; Zhang et al. 1995; 
Hurkman and Tanaka 1996; Hurkman and Tanaka 1996). This enzyme is of interest because 
Cryphonectria parasitica, the chestnut blight fungus, produces large amounts of oxalate at the canker 
margin, which helps lower the pH to toxic levels and binds calcium (Roane et al. 1986). Oxalic acid, or 
oxalate, has also been associated with pathogenesis in other fungi (Noyes and Hancock 1981 ; Marciano et 
al. 1983; Cessna et al. 2000). Therefore, it is reasonable to h}pothesize that neutralizing oxalic acid w ill 
enhance resistance to these fungi. One byproduct of the oxalate degradation. H^O:. could induce a 
separate defense mechanism, which would only be activated when the fungus produces oxalate. H^O: has 
been shown to be a signal molecule that induces a plant's natural defense system (Lane et al. 1993; Lane 
1994) and could enhance resistance in transgenic plants (Wu et al. 1995). PreviousK. we had cloned the 
wheat oxalate oxidase transgene into a model tree species, hybrid poplar, and shown that it could enhance 
resistance to another oxalate producing pathogen. Scploria mmiva (Liang et al. 2001 ). Other researchers 
have cloned this same gene into soybean and shown enhanced resistance to the white mold fungus, 
Sclerotinia scleral iorum (Cober et al. 2003). Recently, transgenic callus from American chestnut 
expressing the oxalate oxidase transgene and grown in the presence of oxalic acid, was shown to retain its 
ability to produce lignin at normal levels. Oxalic acid would signillcantl\ inhibit lignin formation in the 
non-transformed controls (Welch. Slipanovic, Ma\nard. and Powell, unpublished). Lignin s\iilhesis is 
necessaiy to compartmentalize fungal infections in plants. Therefore, since the wheat oxalate oxidase 



80 



gene will have multiple resistance enhancing effects, we believe it will likely enhance blight resistance in 
American chestnut and be sustainable. 

Attached to the oxalate oxidase transgene is a regulated promoter that can control which tissues in the 
plant can express the gene. In this construct, the promoter from the soybean vegetative storage protein B 
(VSPB) gene was chosen because its expression is induced by sucrose and by wounding and repressed by 
auxins (Mason et al. 1993: DeWald et al. 1994). Its expression pattern therefore is primarily in the stems 
and wound sites, the tissues that can be infected by the chestnut blight, and it is not expected to be 
produced in significant amounts in the nuts. The expression of the oxalate oxidase transgene in pVSPB- 
OxO has been tested in Arabidopsis and shows vascular expression as expected (fig. 2). 

A. B. 




^ -♦ 




Figure 2. Oxalate oxidase assays of non-transformed Arabidopsis (A) and 
Arabidopsis transformed with pVSPB-OxO (B). The dark coloration in the 
vascular tissues seen in B is a positive result. 

In addition to the putative resistance-enhancing gene, there are two other transgenes commonly found in 
binary vectors. The first is the BAR gene (Figueira Filho et al. 1994; Metz et al. 1998), which confers 
specific resistance to phosphinothricin (PPT), the active ingredient in the herbicide Finale®. In plant 
transformations, this gene is used for selection of transformed cells. In an American chestnut restoration 
project, this gene will also have a second useful function. To save as much of the genetic diversity, 
including rare alleles, in the surviving population of American chestnuts, a restoration program would 
need to out-cross to as many of these trees as possible. To do vhis efficiently, the transgenes must convey 
resistance in the hemizygous state, meaning full resistance from a single copy. If a hemizygous tree were 
to be out-crossed to a non-transgenic tree, only half of the resulting seeds would contain the transgenic 
resistance cassette. This is where the herbicide resistance would be useful. Nurserymen could plant all 
the seeds and as they sprout, spray them with Finale®. Only the transgenic trees would survive. This 
would save a lot of labor and money, and would increase the efficiency of a restoration program. This 
would also allow easy identification of the transgenic trees in the field by a technique called spotting. In 
this technique, a small drop of Finale® is spotted on a freshly excised leaf If the plant is transgenic, no 
necrotic spot will appear, but after a few days, the non-transgenic plants will display a necrotic spot on the 
leaf 

Lastly, in pVSPB-OxO there is a gene encoding a green fluorescent protein (Haseloff et al. 1997). When 
illuminated by ultraviolet or blue light, this protein will fluoresce green and can be detected using specific 
filters. This gene has greatly enhanced our ability to optimize the transformation procedure because it 
allows visual identification of the transformed tissues without damaging them. I his gene is currently 
being used to optimize our transformation procedures but may or may not be in the Hnal transgenic tree 
released. Although this gene is harmless, we will gage public acceptance belbre using it in the released 
trees. 



81 



CHESTNUT SOMATIC EMBRYO TRANSFORMATIONS 



This year, 2004, marks one hundred years since the 
discovery of the chestnut blight (Merkei 1906). This is also 
the year in which a method has been developed that can 
consistently produce transgenic American chestnut. Over 
the past fourteen years, foundation research has been 
accomplished in Dr. Maynard's and Dr. Merkie's labs 
(Merkle 1991; Carraway et al. 1997; Xingetal. 1997; Xing 
et al. 1999). This year, two advances have greatly improved 
the transfomiation protocol. The first is the use of GPP to 
identify and follow transformed tissues. The second was to 
add a desiccation step to the transformation procedure 
(Cheng 2003). To date, nine lab members and students have 
been able to transform chestnut embryos. Some of the best 
looking transgenic embryos are beginning to develop (fig. 
3). Currently, the first transgenic lines of embryos are being 
propagated and a portion will be stored cr>'Ogenicall\ in Dr. 
Merkle's lab (Holliday 2000). The remainder will begin the 
process of regeneration into whole plants. 




Figure 3. Example of an American 
chestnut somatic embr\o transformed with 
pVSPB-OxO expressing GFP (transformed 
& photographed b\ Ron Rothrock, Dr. 
Maynard's graduate student). 



NEXT STEPS 

In parallel with the transformation work, we have been conducting studies on propagating American 
chestnut somatic-embryo-derived and apicai-meristem-derived plantlets and acclimatizing those plantlets 
for establishment in the field (LaPierre 2003). The first non-transgenic. micropropagated American 
chestnut plantlets were established in a bare-root nursery in 1997 and were lifted, examined for root 
morphology (fig. 4), and transplanted to the field in 2001 . 



The next step in the process of evaluating the new 
transgenic somatic embryos will be to first multiply 
and germinate individual embryos (Merkle et al. 1991, 
Carraway et al. 1997; Xing et al. 1999), or if 
germination is low. micropropagate them (Xing et al. 
1997). Once a sufficient supply of transgenic plantlets 
has been produced, they will be acclimatized and 
grown in the greenhouse (Bickel et al 2000) until they 
have a minimum stem caliper of 3 mm and then 
screened for transgene expression in the stem tissues. 
The trees that express the transgene as expected will 
then be tested for blight resistance. Those 
transformation events producing trees with a high level 
of blight resistance will be planted in field trials and 
evaluated against non-transgenic American and 
Chinese chestnut {Castanca moUissima) seedlings for 
resistance, growth characteristics, and m\conhizal 
interactions. 




Figure 4. American chestnut tissue-culture- 
derived trees (left of meter stick) and seedling 
controls (right of meter stick) after four growing 
seasons. 



82 



CONCLUSIONS 

If all goes according to our projected timeline, we expect to have the first potted transgenic American 
chestnut trees in the greenhouse by the spring of 2005. At this time we will start the regulatory process 
for approved release, beginning with USDA notification of the field trials. These plants will then be 
hardened off and should be ready for field planting in the fall of 2005. We hope that some of the field 
trials can be set up as public educational demonstrations. These plantings will be in controlled areas, but 
will be accessible to the public so that they can observe the ongoing experiments and learn about the 
chestnut blight and about possible uses of forest biotechnology. The final transgene make-up of the 
transgenic American chestnut that will be released to the public will depend on the results from the 
resistance assays, field trials, regulatory approval, and public acceptance. Once approved for release, we 
believe that transgenic American chestnuts trees will play a key roll in the restoration of this threatened 
species. 



ACKNOWLEDGEMENTS 

We would like to thank The American Chestnut Foundation, ArborGen LLC, and the New York State 
American Chestnut Research and Restoration Project for their support of this research. 



LITERATURE CITED 

Bickel, S.L., S.P. LaPierre, W.A. Powell, and C.A. Maynard. 2000. Development of potting mixes for 
acclimatization of American chestnut (Castanea dentata (Marsh. (Borkh)) plantlets. Congress on In Vitro 
Biology. 2000 Meeting of the Society for In Vitro Biology. San Diego, CA, June 10-15, 2000. 
Addendum Book Abstract # P-0032. 

Carraway, D.T. and S.A. Merkle. 1997. Plantlet regeneration from somatic embryos of American 
chestnut. Can. J. For. Res. 27: 1 805- 1812. 

Cessna, S.G., V.E. Sears, M.B. Dickman, and P.S. Low. 2000. Oxalic acid, a pathogenicity factor for 
Sclerotinia sclerotiorimu suppresses the oxidative burst of the host plant. Plant Cell 12(1 1 ):2 19 1-2 199. 

Cheng. M.. T. Hu, J. Layton, C-N. Liu, J.E. Fry. 2003. Desiccation of plant tissues post-Agrobacterium 
infection enhances T-DNA delivery and increases stable transformation efficiency in wheat. In Vitro Cell. 
Dev. Biol. Plant 39:595-604. 

Cober, E.R., S. Rioux, I. Rajcan, P. A. Donaldson, and D.H. Simmonds. 2003. Partial resistance to white 
mold in a transgenic soybean line. Crop Sci. 43( 1 ):92-95. 

Connors, B.J., N.P. Laun, C.A. Maynard. and W.A. Powell. 2002. Molecular characterization of a gene 
encoding a cystatin expressed in the stems of American chestnut {Castanea dentata). Planta 2 1 5(3 ):5 1 0- 

514. 

Connors, B.J., M. Miller, C.A. Maynard, and W.A. Powell. 2002. Cloning and characterization of 
promoters from American chestnut capable of directing reporter gene expression in transgenic 
Arabidopsis plants. Plant Sci. 163(4):77 1-781. 

DeWald, D.B., A. Sadka, and J.E. Mullet. 1994. Sucrose modulation of soybean Vsp gene expression is 
inhibited by auxin. Plant Physiol. 104(2):439-444. 



83 



Dumas, B., G. Freyssinct, and K.E. Pallett. 1995. Tissue-specific expression of germin-like oxalate 
oxidase during development and fungal infection of barley seedlings. Plant Physiol. 107(4): 1091- 1096. 

Figueira Filho, E.S., L.F.A. Figueiredo, and D.C. Monte-Neshich. 1994. Transformation of potato 
{Solarium tuberosum) cv. Mantiqueira us\ng Agrohacterium lumefaciens and evaluation of herbicide 
resistance. Plant Cell Reports l3(12):666-670. 

Haseloff, J., K.R. Siemering, D.C. Prasher, and S. Hodge. 1997. Removal of a crvptic intron and 
subcellular localization of green fluorescent protein are required to mark transgenic ArabiJopsis plants 
brightly. Proc. Nat. Acad. Sci. USA 94(6):2 1 22-2 1 27. ^^ ■ 

Holliday, C.P., and S.A. Merkle. 2000. Preservation of American chestnut germplasm by crvostorage of 
embryogenic cultures. J. Am. Chestnut Found. 14( 1 ):46-52. 

Hurkman, W.J., and C.K. Tanaka. 1996. Effect of salt stress on gemiin gene expression in barley roots. 
Plant Physiol. 11 0(3):97 1 -977. 

Hurkman, W.J., and C.K. Tanaka. 1996. Geniiin gene expression is induced in wheat leaves b\ powdery 
mildew infection. Plant Physiol. 1 1 1(3):735-739. 

Lane. B.C. 1994. Oxalate, germin, and the extracellular matrix of higher plants. J. Fed. Am. Soc. Exper. 
Biol. 8(3):294-30l. 

Lane, B.G., J.M. Dunwell. J. A. Ray, MR. Schmitt, and A.C. Cuming. 1993. Gennin. a protein marker of 
early plant development, is an oxalate oxidase. J. Biol. Chem. 268(1 7): 12239- 12242. 

Lane, B.G., Z.F. Grzelczak, T.D. Kennedy, R. Kajioka. J. Orr, S. D'Agostino, and A. Jaikaran. 1986. 
Germin: compartmentation of two forms of the protein by washing growing wheat embryos. Biochimie et 
Biologic Ccllulaire (Biochemistry and cell biology) 64(10): 1025-1037. 

LaPierre, S. 2003. Studies in American chestnut {Castanea dentata Marsh. (Borkh.)) micropropagation 
and acclimatization. M.Sc. thesis. SUNY College of Environmental Science and Forestry, Syracuse, NY. 

153 p. 

Liang, H., CM. Catranis, C.A. Maynard, and W.A. Powell. 2002. Enhanced resistance to the poplar 
fungal pathogen, Septuria nnisiva. in hybrid poplar clones transfomied with genes encoding antimicrobial 
peptides. Biotechnology Letters 24(5):383-389. 

Liang, H., C.A. Maynard, R.D. Allen, and W.A. Powell. 2001. Increased Septoria musiva resistance in 
transgenic hybrid poplar leaves expressing a wheat oxalate oxidase gene. Plant Mol. Biol. 
45(6):6 19-629. 

Marciano, P.. P. Di Lenna. and P. Magro. 1983. Oxalic acid, cell wall-degrading enzymes and pH in 
pathogenesis and their significance in the virulence of two Sderotitiia sclerotiorum isolates on sunflower 
Helianthus annuus. Physiol. Plant Path. 22(3):339-345. 

Mason, U.S., D.B. DeWald. and J.E. Mullet. 1993. Idcntillcalion of a nieth\ 1 jasmonate-responsive 
domain in the soybean vspB promoter. Plant Cell 5(3):24l-251. 

Merkel. H.W. 1906. A deadly fungus on the American chestnut. N.Y. Zool. Soc. Ann. Rep. 10:97-103. 



84 



Merkle. S.A., A.T. Wiecko. and B.A. Watson-Pauley. 1991. Somatic embryogenesis in American 
chestnut. Can. J. For. Res. 21:1 698- 1 70 1 . 

Metz, P.L.J., W.J. Stiekema, and J. P. Nap. 1998. A transgene-centered approach to the biosafety of 
transgenic phosphinothricin-tolerant plants. Mol. Breed. 4(4):335-341. 

Noyes. R.D.. and J.G. Hancock. 1981. Role of oxalic acid in Sclerotinia wilt of sunflower. Physiol. Plant 
Path. 18:123-132. 

Powell, W.A., CM. Catranis. and C.A. Maynard. 1995. Synthetic antimicrobial peptide design. 
Molecular plant-microbe interactions: Molecular Plant-Microbe Interactions 8(5):792-794. 

Powell, W.A., CM. Catranis, and C.A. Maynard. 2000. Design of self-processing antimicrobial peptides 
for plant protection. Letters in Applied Microbiology 3 1(2): 163-168. 

Powell. W.A., and C.A. Maynard. 1997. Designing small antimicrobial peptides and their encoding 
genes. Micropropagation, Genetic Engineering, and Molecular Biology of Populus, Fort Collins, CO. 

U:165-172. 

Roane, M.K., G.J. Griffin, and J.R. Elkins. 1986. Chestnut Blight, Other Endothia Diseases, and the 
Genus Endothia. Am. Phytopathol. Soc, APS Press, St. Paul, MN. 

Wu, G., B.J. Shortt, E.B. Lawrence. E.B. Levine, K.C Fitzsimmons, and D.M. Shah. 1995. Disease 
resistance conferred by expression of a gene encoding H202-generating glucose oxidase in transgenic 
potato plants. Plant Cell 7(9): 1357- 1368. 

Xing, Z., W.A. Powell, and C.A. Maynard. 1999. Development and germination of American chestnut 
somatic embryos. Plant Cell, Tissue and Organ Culture 57(1 ):47-55. 

Xing. Z.H., M.F. Satchwell. W.A. Powell, and C.A. Maynard. 1997. Micropropagation of American 
chestnut: increasing rooting rate and preventing shoot-tip necrosis. In Vitro Cellular & Developmental 
Biology Plant: J. Tissue Cuhure Assoc. 33(l):43-48. 

Zhang, Z., D.B. Collinge, and H. Thordal-Christensen. 1995. Germin-like o.xalate oxidase, a 
H202-producing enzyme, accumulates in barley attacked by the powdery mildew fungus. Plant J. Cell 
Mol. Biol. 8(1):139-145. 



85 



86 



Steiner, K. C. and Carlson, J. E, eds. 2006. Restoration of American Chestnut To Forest Lands - Proceedings of a 
Conference and Workshop. May 4-6, 2004, The North CaroHna Arboretum. Natural Resources Report 
NPS/NCR/CUE/NRR - 2006/001, National Park Service. Washinuton. DC. 



H\ POVIRULENCE: USE AND LIMITATIONS AS A CHESTNUT BLIGHT 

BIOLOGICAL CONTROL 

William L. MacDonald and Mark L. Double 

Division of Plant and Soil Sciences, West Virginia University, 

Morgantown, WV 26506-6058 USA (macd@wvu.edu) 



Abstract: The recovery of chestnut from chestnut blight in Italy and Michigan largely was responsible for 
the resurgence in chestnut research. The observed remission of disease now has been attributed to a 
biological control process called hypovirulence, whereby virulent strains are debilitated as a result of 
infection by fungal viruses (hypoviruses). Several species of hypoviruses now are known and each may 
impart unique effects on Cryphonectria parasitica. Lethal infections often are controlled by introducing 
the appropriate hypovirus into cankers. Unfortunately, at many locations within the native range of 
American chestnut, a complex system of vegetative incompatibility restricts hypovirus transmission 
among strains. Factors like vegetative incompatibility apparently regulate the widespread establishment 
of hypoviruses and presumably are, in part, responsible for our inability to artificially establish 
hypoviruses to the extent that has occurred naturally. Some of the factors that regulate hypovirus success 
or failure may be discovered as part of ongoing research at an isolated Wisconsin chestnut stand. 
Hopefully, understanding the phenomenon of hypovirulence eventually will allow it to be employed as 
part of the American chestnut restoration program. 



INTRODUCTION 

Cryphonectria parasitica was first recognized one hundred years ago as the fungus responsible for the 
cankers that resulted in the death of American chestnut (Merkel 1905). This brightly orange-pigmented 
organism was new to North America and by the time it was identified, its role in the tragedy that was 
about to unfold was cast. The remarkably detailed work of several early scientists unraveled the biological 
details of a host-pathogen relationship that would have unparalleled ecological and sociological impact 
(Anderson and Babcock 1913, Heald and Gardner 1913, Shear and Stevens 1913). Within ten years of 
the identification of C parasitica as the causal fungus, their writings sadly were predictive of what was to 
ensue. As the blight spread, most of the research efforts turned to the strategy of breeding blight resistant 
trees (Fleet 1914). The early breeding programs met with limited success and never were designed to 
control blight in eastern forests. For almost fifty years, relatively little research attention was directed 
toward the causal fungus. If this same organism was introduced today, American chestnut undoubtedly 
would face the same fate. 

We are fortunate that American chestnut was saved from extinction in its natural range by its propensity 
to sprout. The first one hundred years of chestnut blight is a blink in biological time, it may, however, be 
this surviving sprout population that over longer biological time periods allows for the expression of a 
disease of C parasitica that may result in natural biological control of chestnut blight. A glimmer of 
hope that this was possible emerged from observations of "spontaneously healing" cankers that were 
noted on European chestnut growing in northern Italy (Grente 1965). The "recover)/" phenomenon was 
confirmed when Jean Grente, a French mycologist, described a variety of unusual strains of C. parasitica 
that he isolated from the callusing cankers (Grente and Berthelay-Sauret 1978). The isolates he recovered 
often were lightly pigmented in contrast to the normal orange pigmented lethal strains. Moreover, Grente 
found that, like the infection from which they were isolated, when the isolates were inoculated to healthy 



87 



chestnuts, lethal infections seldom resulted. Most significant was the observation that the cause of this 
debilitation was transmissible to virulent strains. Grente coined the term "hvpovirulent" to describe the 
reduced state of virulence and suggested that "cytoplasmic agents" were responsible for this phenomenon 
(Grente 1965). Remarkably, for the chestnut growing in northern ItaK, this marked ""reco\er\"" was 
occurring within twenty-five years of the discovery of chestnut blight in Europe (Mittempergher 1978). 

By the time recovery was detected in Italy, the spread of the chestnut blight fungus through the natural 
range of American chestnut was complete. There were few, if any, signs of resistance or recoven. from 
infection. Grente's discover)' and his research describing hypovirulence refocused attention on chestnut 
blight in this country, especially in the laboratories at The Connecticut Agricultural Experiment Station 
where a longstanding chestnut breeding program still was active. Further, the phenomenon of 
hypovirulence brought attention to the pathogen. It also was during this period of the early 1970s that a 
small stand of blighted chestnuts growing in Michigan was brought to the attention of the Connecticut 
research team (Elliston et al. 1977). They quickly discovered that the isolates from Michigan also carried 
a "cytoplasmic agent" that was transmissible and reduced the ability ofC. parasiiica to produce lethal 
cankers (Elliston 1985). Unlike the hypovirulent strains from Italy, the Michigan isolates retained their 
bright orange pigmentation. As the search for chestnut trees that were recovering from blight in Michigan 
intensified, more and more "recovering stands" were discovered, all outside the natural range of chestnut. 
Although chestnut blight is still the dominant stressing agent in most of the isolated Michigan stands, in 
some, the impact of the disease is minor, and the level of recover) mimics the expression of 
hypovirulence in some areas of Italy (Fulbright et al. 1983). 

Since these two remarkable settings have been described, hypovirulent strains have been identified at 
various locations within the natural range of American chestnut and some are associated with surviving 
trees (GrifTin 2000). Unfortunately, the wide-spread recover) of chestnut as a result of the hypovirulence 
phenomenon is unknown in areas where sprout populations still persist. Several reviews of the recover) 
of chestnut blight and associated hypovirulence are published (Milgroom and Cortesi 2004, MacDonald 
and Fulbright 1991, Heiniger and Rigling 1994, Nuss 1992, Van Alfen 1992). 



WHAT IS HYPOVIRULENCE? 

A variety of factors control the level of virulence in C. parasitica including the genetic makeup of the 
fungus or a variety of nonviral, cytoplasmic agents, such as defective mitochondria or plasmids. 
However, the term hypovirulence most often refers to the reduction in virulence caused by fungal viruses. 
The first indication that virus-like agents might be involved came with the association of double-stranded 
(ds) RNA with the European and North American strains that were shown to be less \ irulent (Da) et al. 
1977). These dsRNAs eventually were shown to represent a unique group of viruses, now called 
hypoviruses (Hillman et al. 2000). The definitive proof of the cause and effect relationship and their 
infectious nature occurred through the application of molecular technology (Choi and Nuss 1992). 
Although fungal-virus associations have been know for decades, the h)poviruses associated with C. 
parasitica are unique; rather than being encapsulated in a protein coat, the) are membrane bounded 
(Newhouse et al. 1983). As a result, a new \ irus famil). the Hypoviridac. has been established for the 
four species (CHV 1 though CHV4) of hypoviruses that have been discovered to date (Hillman and Suzuki 
2004). Most studies have been of the CHVl species, as it was the first hypovirus identified and is the 
hypovirus associated with biological control of chestnut blight in Europe (Shapira et al. 1991). This 
hypovirus also has been discovered infecting strains of C. parasitica in China and Japan, but it has never 
been identified as a natural component of ('. parasitica in North America (Pee\er et al. 1998). The 
CFW3 hypovirus is associated with the recovering Michigan chesiiuit stand?> but its origin remains 
unknown as it has not been isolated in the Orient (l^aul and Fulbright 1988). CHV2 is uncommon and 
known only from a site in New Jersey (Hillman et al. 1994). It also has been identified in C. parasitica 



88 



populations in China (Peever et al. 1998). CHV4 is somewhat unique; unhke CHV1-CHV3. it has little 
or no observable effect on the virulence or other traits of C. parasitica (Enebak et al. 1 994). It is 
widespread in its association with isolates from the central Appalachians but its origin and role remain 
undiscovered. 

The effects of hypovirus infection on the blight fungus are variable and appear to be a function of the C. 
parasitica strain as well as the infecting hypovirus (Chen and Nuss 1999, MacDonald and Double 1998) 
For those hypoviruses that reduce fungal virulence, infection often results in smaller non-lethal cankers 
and a corresponding reduction in the production of asexual spores and almost certainly the reduction or 
elimination of sexual sporulation. What currently is known about the molecular influence of the 
hypovirus on the physiological processes of the fungus has been reviewed (Nuss 1992, Nuss 1996). 



EXPLOITING HYPOVIRULENCE 

The discovery of hypovirulence and the observation of a notable level of disease control on American 
chestnut in Michigan brought hopes for the first time that some level of biological control was possible in 
North America. Procedures first employed by Grente to treat virulent infections were duplicated. 
Subsequently, modifications to Grente's treatment protocols and a variety of different inoculum types 
were used to introduce hypoviruses into virulent cankers on American chestnut sprouts (Hobbins et al. 
1992, MacDonald and Double 1979). The results often were ver> encouraging as hypovirus transfer 
frequently occurred and the expansion of individual treated infections frequently was arrested as callus 
tissue formed at the margins of cankers. Even though many of the treatments were successful and the life 
of sprouts was prolonged, the sheer number of subsequent infections that developed on the same stem 
dramatically weakened the tree, and when some cankers were not arrested by treatment, trees died 
(MacDonald and Fulbright 1998). Further, there was little evidence that natural hypovirus spread on the 
same stem afforded any protection to other virulent infections that almost certainly would arise. With few 
exceptions, most hypovirulent introductions were unsuccessful if measured by the number of treated 
sprouts that remained alive several years after treatment (Milgroom and Cortesi 2004)). 

As a result of the early releases, several factors were discovered that may influence the effectiveness of 
the hypovirulent treatments. When additional hypovirulent strains were discovered and their infecting 
hypoviruses investigated, the variation in their effects on C. parasitica became apparent. Some virulent 
strains were so debilitated by hypovirus infection that they grew poorly in bark and almost completely 
failed to produce hypovirulent inoculum (Double and MacDonald 1995). Therefore, concern arose that 
highly debilitating hypoviruses have such an extreme effect on their fungus host that there is little 
potential for the strains to grow in bark and produce inoculum to perpetuate themselves. A sense 
developed that hypoviruses that do not debilitate C. parasitica as significantly may be more useful 
biological control agents (MacDonald and Fulbright 1998). Logically, if strains are more capable of 
invading bark and generating hypovirulent inoculum without killing their hosts, they may be more 
capable of disseminating their hypoviruses and thus potentially better biological control agents. 



ROLE OF VEGETATIVE COMPATIBILITY 

Early laboratory and field experimentation also revealed that an incompatibility system existed in C. 
parasitica (Anagnostakis 1977). When strains are incompatible, their hyphal elements fail to fuse 
(anastomose), restricting cyptoplasmic and hypovirus exchange. Unlike many plant and animal viruses, 
viruses that infect fungi have no extracellular phase and therefore must be transmilled to progeny in 
spores during reproduction or via hyphal anastomosis and cytoplasmic mixing. The system of vegetative 
compatibility in C. parasitica, as in other fungi, represents a self-recognition system that prevents 



89 



incompatible strains from fusing. Essentially, as the hyphal filaments of the fungus approach each other, 
cell death occurs, restricting the fusions necessary for hypovirus transmission. The system of vegetative 
incompatibility in C pcinisltica is controlled by at least six genes (Huber 1996. Cortesi and Milgroom 
1998). fhe probabilit} of hypovirus transmission is high when strains share identical genes. 
Transmission is less likely if gene differences exist with probabilities of transmission related to the 
number of gene differences and the specific genes present. 

Considerable research on vegetative compatibilitv has been conducted (Cortesi and Milgroom 1998. 
Milgroom 1995, Milgroom and Cortesi 1999). One interesting relationship that has been discovered 
relates to the diversity of vegetative compatibility tvpes and hypovirus transmission in tleld settings. In 
general, sites where biological control generally is more successful have a less diverse population of 
vegetative incompatibility genes (Milgroom et al. 1996). This appears to be the situation in Italy and 
Michigan (Cortesi et al. 1996) where the number of vegetative compatibilit>' genes that are expressed is 
quite low when compared to the diversity that occurs in Asia or the central Appalachian region (Figure 1). 
Whether the lack of diversity is responsible for the widespread distribution of hvpoviruses and biological 
control that has occurred is unknown. 

Table 1 . Diversity of vegetative compatibility types in four chestnut areas (two in the U.S. from 
Castanea dentata stands, one in Italy from a C. sativa stand, one in China from a C. niollissima stand and 
one in Japan from a C. crenata stand). 



Population 


Number of isolates tested 


Number of VC tvpes 


Finzel, MD 


57 


25 


Bartow, WV 


61 


29 


Italy* 


716 


20 


China* 


79 


71 


Japan* 


30 


29 



*Data from Milgroom 

A second interesting relationship between hypovirus infection and the diversit\ of vegetative 
compatibility t>'pes is the effect hypovirus infection mav have on the diversity of vegetative compatibilit\' 
types at a site. Sexual reproduction is responsible for maintaining diversity (Marra and Milgroom 2001 ). 
One must therefore consider whether the low diversity that exists at some recovering sites is because 
hypovirus infections have reduced sexual reproduction or the reduced diversity has been a longstanding 
feature of the stand and has permitted hypoviruses to be disseminated successfully. 

Although vegetative compatibility diversity appears to influence the success of biological control, other 
factors also may be involved. Certainly the role of the host cannot be overlooked in the expression of the 
hypovirulence phenomenon. In tests of susceptibility, European chestnuts consistently have been shown 
to be slightly more resistant than American chestnut (Bazzigher 1981 ). A more resistant host almost 
certainly would provide a longer time period for infections to acquire hv po\ iruses. perhaps enough time 
for the successful expression of the hvpov irulence. Similarly, the ecosv stems in which the two species 
typically grow are quite different. In its North American range, chestnut grows among a diverse mix of 
competing hardwoods. This often is not the case in Europe, especially in areas where European chestnut 
is cultivated for nut or coppice production (Bounous 1999). Likewise, at many recovering sites in 
Michigan, chestnuts grow in the absence of signitlcant competition from other species. These settings 
permit the constant regeneration of chestnut biomass and mav in turn foster the dissemination of 
hypoviruses. Unfortunately, in eastern North America, chestnut is largelv a relic in the understory with 
little opportunit> to grow and develop significant numbers of canker to even acquire hypoviruses. One 



90 



site where many of these limitations do not exist is in a stand of American chestnut growing near West 
Salem, Wisconsin. 



UTILIZING HYPOVIRUSES AT WEST SALEM 

The West Salem stand is the largest stand of American chestnut in the U.S. The origin of the stand dates 
to the late 1 800s when a few chestnuts were planted at the site by the landowner who had moved there 
from the eastern U.S. (Cummings Carlson et al. 2002). Chestnut now is the dominant species on about 50 
acres of land. Unfortunately, in 1986, C. panisilica was discovered at the site and now threatens the 
future of this magnificent stand. Attempts from 1988-90 to eradicate the fungus failed. A biological 
control program was initiated in 1992. At that time, the stand seemed to present an excellent opportunity 
to exploit hypoviruses for two reasons. First, there were few trees infected so the disease was at the very 
early stage of the epidemic. Second, the stand was infected by a single clone of C. parasitica: hence, the 
barriers imposed by vegetative compatibility did not exist (McGuire and Milgroom 2002). Over the next 
six years, two hypoviruses were introduced into the resident West Salem strain (Double and Mac Donald 
2002). These were deployed by introducing inoculum into small holes made around the margin of the 
canker. The first hypovirus (CHV3 type) deployed (1992-94) was obtained from a recovering grove of 
chestnuts near Cadillac, Michigan. The second was a hypovirus (CHVl type) associated with an Italian 
hypovirulent strain and was used for treatment from 1995-97. As cankers were treated, they routinely 
acquired hypovirus and the treated chestnuts responded by producing significant callus growth to close 
the infection. 

Between 1992-1997, about 650 cankers on 138 trees were treated. To assess hypovirus spread each 
season, 8-12 small bark plugs were removed from the treated cankers and also from new cankers that had 
formed. An evaluation of hypovirus infection was made when the plugs were cultured and the cultural 
appearance of the resulting colonies was compared to that of virulent strains. 

From 1998-2002, no additional hypovirus introductions were made, so that an evaluation of the level of 
natural spread could be made over several seasons. The results have been mixed. Treated cankers have 
retained hypoviruses and many are heavily callused and blight is no longer damaging (Table 1 ). 
Likewise, new cankers that have developed on trees with treated cankers have acquired hypovirus at high 
levels. Many of these stems are still alive almost ten years after initial treatment. Unfortunately, 
hypoviruses have not spread significantly to cankers on nearby trees that never received hypovirus 
treatment (Double and MacDonald 2002). Because the number of new virulent infections continues to 
increase rapidly, the decision was made in 2003 to once again introduce hypoviruses. If biological 
control can be initiated on individual trees by canker treatment, as seems to be indicated, the additional 
treatments may help save some trees and also help determine why viruses do not disperse to new cankers 
on trees that are untreated. 



91 



Table 2. Classification of cankers at West Salem, WI based on cultural morphology of Cryphonectria 
parasitica isolates removed from treated cankers, non-treated cankers on treated trees and non-treated 
cankers on non-treated trees from 1994-2000. 





Hypovirus-treated 
cankers 


Non-treated cankers on 
treated trees 


Non-treated cankers on 
non-treated trees 


Year Sampled 


Hypovirulent 


Virulent 


llypovirulent 


Virulent 


H\po\ irulent 


Virulent 


1994 


55% 


45% 


18% 


82% 


29% 


71% 


1995 


55% 


45% 


22% 


78% 


()"o 


lOO-'o 


1996 


80% 


20% 


58% 


42%, 


33% 


67% 


1997 


82% 


18% 


42% 


58% 


9% 


91% 


1998 


83% 


17% 


71% 


29% 


22% 


78% 


1999 


81% 


19%, 


71% 


29% 


28% 


72% 


2000 


60% 


40% 


46% 


54% 


9% 


91% 



CONCLUSIONS 

After its initial discover^', the prospects of utilizing hypoviruses to biologically control chestnut blight 
seemed reasonably straightforward and exploitable. The early successes achieved by treating infections 
on stems were in themselves remarkable. Unfortunately, hypovirus populations have not been 
perpetuated or disseminated adequately at sites where they have been artificially released, as has 
happened naturally at some locations. Admittedly, the phenomenon of hypovirulence. like most 
biological issues, is wrought with complexity. Whether we can duplicate artificially what has happened 
naturally remains a significant question. Clearly, major details of the epidemiologv of chestnut blight and 
the infecting hypoviruses need to be unraveled. The successful transition of a \ irulent C. parasitica to 
one that is laden with debilitating hypoviruses is not regulated by a single factor. .A summar\ of some of 
the components involved in the expression of hypovirulence that require further research include: 

• an understanding of the contribution of chestnut genotvpe to the expression of hypovirulence; 

• an appreciation of the role environmental factors play in contributing to the success or failure of 
hypovirulence; 

• an evaluation of the pathogen population relative to its ability to cause disease, reproduce and acquire 
hypoviruses; 

• an assessment of the intluence specific hypovirus species have on C. parasitica; 

• a consideration of potential vector relationships that might influence hypovirus dissemination; and, 

• an evaluation of strategies for the deployment of hypoviruses. 

We remain encouraged by the levels of recovery from blight thai has occurred naturall\ at some locations. 
Over long biological time periods, hypovirulence ma\ emerge on its own within the American chestnut's 
natural range. But, if man is to inlluence the process of biological control, a more complete 
understanding of the biology of the hypovirulence phenomenon is required. 



92 



LITERATURE CITED 

Anagnostakis, S.L. 1977. Vegetative incompatibility in Endothia parasitica. Exper. Mycol. 1 :306-316. 

Anderson, P. J., and D.C. Babcock. 1913. Field studies on the dissemination and growth of the chestnut 
tree fungus. Penn. Chestnut Tree Blight Commission Bull. 3:1-32. 

Bazzigher, G. 1981. Selection of blight-resistant chestnut trees in Switzerland. Eur. J. For. Path. 1 1:199- 
207. 

Bounous, G. 1999. Among the chestnut trees in Cuneo Province. Edizione Metafore, Cuneo, Italy. 79p. 

Chen, B., and D.L. Nuss. 1999. Infectious cDNA clone of a hypovirus CHVl-Euro7: A comparative 
virology approach to investigate virus-mediated hypovirulence of the chestnut blight fungus 
Cryphonectria parasitica. J. Virology 73:985-992. 

Choi, G.H., and D.L. Nuss. 1992. Hypovirulence of chestnut blight fungus conferred by an infectious 
viral cDNA. Science 257:800-803. 

Cortesi, P., M.G. Milgroom, and M. Bisiach. 1996. Distribution and diversity of vegetative compatibility 
types in subpopulations o^ Cryphonectria parasitica in Italy. Mycol. Res. 100:1087-1093. 

Cortesi, P.. and M.G. Milgroom. 1998. Genetics of vegetative incompatibility in Cryphonectria 
parasitica. Appl. Environ. Microbiol. 64:2988-2994. 

Cummings Carlson, J., F.L. Paillet, and S.E. Dahir. 2002. Ecological history and early disease 
management of an isolated stand of American chestnut in Wisconsin. Phytopathology 92:S94. 

Day, P.R., J. A. Dodds, J.E. Elliston, R.A. Jaynes, and S.L. Anagnostakis. 1977. Double-stranded RNA in 
Endothia parasitica. Phytopathology 67: 1 393- 1 396. 

Double, M.L.. and W.L. MacDonald. 1995. The influence of month of inoculation with development of 
Cryphonectria parasitica cankers in American chestnut stems. P. 481-484 /// International Congress on 
Chestnut, October, 1993, Antognozzi, E. (ed.). Spoleto, Italy. 

Double, M.L., and W.L. MacDonald. 2002. Hypovirus deployment, establishment and spread: results 
after six years of canker treatment. Phytopathology 92:S94. 

Elliston, J.E. 1985. Preliminary evidence for two debilitating cytoplasmic agents in a strain of Endothia 
parasitica from western Michigan. Phytopathology 75:170-173. 

Elliston, J.E., R.A. Jaynes, PR. Day, and S.L. Anagnostakis. 1977. A native American hypovirulent 
strain of Endothia parasitica. Proc. Am. Phytopath. Soc. 4:83. 

Enebak, S.A., W.L. MacDonald, and B.I. Hillman. 1994. Effect of dsRNA associated with isolates of 
Cryphonectria parasitica from the central Appalachians and their relatedness to other dsRNAs form 
North America and Europe. Phytopathology 84:528-534. 

Fleet, W. Van. 1914. Chestnut breeding experience. J. Hered. 5:19-25. 



93 



Fulbright D.W.. W.H. Weidlich, K.Z. Haufler, C.S. Thomas, and C.P. Paul. 1983. Chestnut blight and 
recovering American chestnut trees in Michigan. Can. J. Bot. 61:3164-3171. .^ 

Grente. J. 1965. Les formes hypovirulentes d"Endothia parasitica et les espoirs de lutte contre le chancre 
du chantaingnier. C.R. Acad. Agric. France 51:1033-1037. 

Grente, J., and S. Berthelay-Sauret. 1978. Biological control of chestnut blight in France. P. 30-34 in 
Proceedings of the American Chestnut Symposium, W.L. MacDonald. F.C. Cech. J. Luchok and C. 
Smith, (eds.). West Virginia Universitv Press. 

Griffin, G.J. 2000. Blight control and restoration of the American chestnut. J. For. 98:22-27. ,^ 

Heald, F.D., and M.W. Gardner. 1913. The relative prevalence of pycnospores and ascospores of the 
chestnut blight fungus during the winter. Phytopathology 3:296-305. 

Heiniger, U., and D. Rigling. 1994. Biological control of chestnut blight in Europe. Ann. Rev. Phytopath. 
32:581-599. 

Hillman, B.I.. B.T. Halpem, and M.P. Brown. 1994. A viral dsRNA element of the chestnut blight fungus 
with a distinct genetic organization. Virology 201 :24 1-250. 

Hillman. B.I.. D.W. Fulbright, D.L. Nuss, and N.K. VanAlfen. 2000. P. 515-520 in: Hypoviridae: In 
Virus Taxonomy: Seventh Report of the International Committee for the Taxonomy of Viruses. M.H.V.v 
Regenmortel, CM. Fauquet. D.H.L. Bishop, E.B. Carstens. and M.K. Estes (eds.). San Diego Academic 
Press. 

Hillman. B.I., and N. Suzuki. 2004. Viruses of the chestnut blight fungus, Cryphonectria parasitica. 
Adv. Virus Res. (in press). 

Hobbins, D.L., M.L. Double, C.R. Sypolt, and W.L. MacDonald. 1992. Interactions between artificially 
established virulent Cryphonectria parasitica cankers and sources of virulent and h\po\ irulent inoculum 
on American chestnut stems. P. 156-161 in Proceedings of the International Chestnut Conference, 
Double, M.L., and W.L. MacDonald (eds.). West Virginia Universit> Press. 

Huber, D.H. 1996. Genetic analysis of vegetative incompatibility pol\niorphisms and horizontal 
transmission in the chestnut blight fungus Cryphonectria parasitica. Ph.D. thesis, Michigan State 
University, East Lansing. MI. 

MacDonald, W.L., and M.L. Double. 1979. Effectiveness of slurry treatments in controlling individual 
Endothia parasitica cankers on American chestnut. Smith, H.C. (ed.). USDA Gen. Tech. Rep. NE-64. 

MacDonald, W.L.. and M.L. Double. 1998. Variation in growth and sporulation of Cryphonectria 
parasitica isolates as influenced by hypovirus infection. Ihe Second International Chestnut Symposium, 
October, 1998, Salesses, G. (ed.). Bordeaux, France. 

MacDonald, W.L., and D.W. Fulbright. 1 991. The biological control of chestnut blight: use and 
limitations of transmissible hypovirulence. Plant Dis. 75:656-661. 

Marra. R.E.. and M.G. Milgroom. 2001. fhe mating system of the fungus, Cryphonectria parasitica: 
selfing and self-incompatibility. F^eredit> 86:134-143. 



94 



McGuire, I.C, and M.G. Milgroom. 2002. Clonal population and reproductive biology of Cryphoneclria 
parasitica. Phytopathology 92:S94. 

Merkel, H.W. 1905. A deadly fungus on the American chestnut. N.Y. Zool. Soc. Ann. Rep. 10:97-103. 

Milgroom, M.G. 1995. Population biology of the chestnut blight fungus, Ciyphonectria parasitica. Can. 
J. Bot. 73:S311-S319. 

Milgroom, M.G., and P. Cortesi. 1999. Analysis of population structure of the chestnut blight fungus 
based on vegetative incompatibility genotypes. Proc. Natl. Acad. Sci. USA 96:10518-10523. 

Milgroom, M.G., and P. Cortesi. 2004. Biological control of chestnut blight with hypovirulence: a critical 
analysis. Ann. Rev. Phytopath. (in press). 

Milgroom, M.G., K. Wang, Y. Zhou. S.E. Lipari, and S. Kaneko. 1996. Intercontinental population 
structure of the chestnut blight fungus, Cryphonectria parasitica. Mycologia 88: 1 79-190. 

Mittempergher, L. 1978. The present status of chestnut blight in Italy. P. 34-37 in Proceedings of the 
American Chestnut Symposium, W.L. MacDonald, F.C. Cech, J. Luchok, and C. Smith, (eds.). West 
Virginia University Press. 

Newhouse, J.R., H.C. Hock, and W.L. MacDonald. 1983. The ultrastructure of Endothia parasitica. 
Comparison of a virulent with a hypovirulent strain. Can. J. Bot. 61 :389-399. 

Nuss, D.L. 1992. Biological control of chestnut blight: An example of virus-mediated attenuation of 
fungal pathogenesis. Microbiol. Rev. 56:561-576. 

Nuss, D.L. 1996. Using hypoviruses to probe and perturb signal transduction processes underlying fungal 
pathogenesis. Plant Cell 8:1845-1853. 

Paul, C.P.. and D.W. Fulbright. 1988. Double-stranded RNA molecules from Michigan hypovirulent 
isolates of Enc/othia parasitica vary in size and sequence homology. Phytopathology 78:751-755. 

Peever, T.L., Y-C Liu, K. Wang, B.l. Hillman, R. Foglia, and M.G. Milgroom. 1998 Incidence and 
diversity of double-stranded RNAs infecting the chestnut blight fungus, Cryphonectria parasitica, in 
China and Japan. Phytopathology 88:81 1-817. 

Shapira, R., G.H. Choi, and D.L. Nuss. 1991. Virus-like genetic organization and expression strategy for 
a double-stranded RNA genetic element with biological control of chestnut blight. EMBO J. 10:731-739. 

Shear, C.L.. and N.E. Stevens. 1913. The chestnut blight parasite {Endothia parasitica) from China. 
Science 38:295-297. 

Van Alfen, N.K. 1982. Biology and potential for disease control of hypovirulence of Endothia parasitica. 
Ann. Rev. Phytopath. 20:349-362. 



95 



96 



Steiner, K. C. and Carlson, J. E, eds. 2006. Restoration of American Chestnut To Forest Lands - Proceedings of a 
Conference and Workshop. May 4-6, 2004, North Carolina Arboretum. Natural Resources Report 
NPS/'NCR/CUE/NRR - 2006/001, National Park Service. Washington, DC. 

INTEGRATED USE OF RESISTANCE, HYPOVIRULENCE, AND FOREST 

MANAGEMENT TO CONTROL BLIGHT ON AMERICAN CHESTNUT 

" ■^.. - ' , 
"~ Gary J. Griffin' ", John R. Elkins", Dave McCurdy"\ and S. Lucille Griffin- 

Department of Plant Pathology, Physiology and Weed Science, 
Virginia Tech, Blacksburg, VA 24061 USA (gagriffi@vt.edu) 
'American Chestnut Cooperators' Foundation, Beaver, WV 25813, USA, and Newport, VA 24128 USA 
West Virginia Division of Forestry, West Columbia. WV 25287 USA 



Abstract: In the natural range of the species, the survival of most large American chestnut trees was 
associated with blight resistance, hypovirulence (reduced virulence) in the chestnut blight fungus, 
Cryphonectria parasitica, and favorable sites. Controlled intercrosses of these trees have resulted in 
progeny with acceptable levels of resistance, which have been used in further breeding. Some American 
chestnuts in a very large population (135,123 trees) derived from open pollinations of large survivors 
have shown promising levels of field blight resistance. In forest clearcuts and plantations, hypovirulence, 
associated with hypovirus infection in the blight fungus, develops naturally following blight epidemics. 
However, chestnut stems die due to: I ) the high blight susceptibility of American chestnut; 2) the rapid 
spread of vegetative compatibility-diverse, abundant, virulent inoculum; 3) the slow spread of 
hypovirulence; 4) high hardwood competition; and 5) low-temperature stress at high-altitude (> 2.500 
feet) sites or drought. A long term (>20 years) and high level of blight control has been obtained on 
mesic, managed (control of competing hardwoods) sites, established with blight-resistant American 
chestnuts that were inoculated with a hypovirulent strain mixture. Cultural studies and nucleotide 
sequence analysis of two hypovirus regions (both >800 bp) indicated that blight control was associated 
with the spread of Italian Cryphonectria hypovirus 1 (CHVl ). Blight resistance may allow time for 
CHVl to spread. Mesic shallow coves on lower altitude slopes are among the best sites to implement 
integrated blight management and restoration of American chestnut. 

Key Words: American chestnut / Blight resistance / Hypovirulence / CHVl / Forest management 



INTRODUCTION 

Following the chestnut blight pandemic, only a few timber-sized American chestnuts survived in the 
native range of the species, where it was once a dominant component of the former oak-chestnut and 
mixed mesophytic forest regions (Braun 1950). From the mid 1970s to the early 1980s, we investigated 
why these trees survived. The main focus was on tests for blight resistance, hypovirulence in the chestnut 
blight fungus, Cryphonectria parasitica (Murrill) Barr, and site factors. Several large trees were found, 
with the largest being greater than 40 inches in diameter at breast height (dbh). This tree grows at the 
base of the Northern Blue Ridge Mountains in Virginia. Understanding why these trees survived may aid 
in attempts to control the blight and restore the American chestnut. 



EVALUATION OF BLIGHT RESISTANCE AND HYPOVIRULENCE IN LARGE, SURVIVING 

AMERICAN CHESTNUT TREES 

The concept of disease resistance in plants, as viewed by American chestnut workers, both scientists and 
non-scientists, is often very restricted. Some believe a plant almost has to be immune to be labeled 
"resistanf to a given disease and that resistant plants never die from disease. Scientists working in the 



97 



field of disease resistance recognize that disease resistance is variable. For example, the grey scale shown 
in F^ig. 1 shows how disease severity can v ar\ from a ver>' low level (represented b\ white) through 
increasing intensities of grey until the scale is black, representing high disease se\erit\'. These disease 
severity levels are then translated into disease resistance/susceptibility ratings, which, therefore, can vary 
from high disease resistance to high disease susceptibilit> . Some workers, such as disease-resistance 
scholar. J.E. Vanderplank (1982), have called the high disease resistance in Fig. 1 "full" or "complete 
resistance", and the high susceptibility "full" or "complete susceptibilit>". All in-between values were 
called "partial resistance" by Vanderplank. When chestnut stems of a different species or genotvpe were 
inoculated with a virulent strain of the chestnut blight fungus, disease severity ratings were continuously 
variable (Anagnostakis 1992: Bazzigher and Schmid 1962: Clapper 1952: Griffin et al. 1983: Hebard 
1999), and thus resistance/susceptibility ratings can vary similarly, as represented in Fig. 1. The middle 
of the scale (see arrow) may be called partial resistance or moderate resistance (MR). The disease severity 
rating can be influenced by environmental factors, ontogenic factors, virulence of the pathogen, and the 
time interval over which the test is conducted (Bazziger and Schmid 1962: Griffin et al. 1986). Stressful 
environmental factors shown to be associated with more severe chestnut blight, such as early frosts (Berry 
1951), growth in frost pockets or at high altitude (Headland et al.l976: Jones et al. 1980), drought (Goa 
and Shain 1995), and low light intensity (Uchida 1977), may increase disease severity (lower arrow by the 
scale). Factors favorable to the chestnut tree, such as optimal soil moisture (Goa and Shain 1995) over 
the growing season, a mesic site, and high intensity light (Uchida 1977). may lower disease se\erit> 
(upper arrow by the scale) ratings (Fig. 1 ). In the former oak-chestnut region, low temperatures at high 
altitudes and summer droughts are common events and can be stressful to chestnuts during blight 
resistance trials. 

Several approaches were used by us to evaluate blight resistance in large, sur\ iving American chestnut 
trees, including /// situ inoculations of branches on the large, sur\ iving trees \\ ith standard \ irulent strains 
of the blight fungus, inoculation of seedling progeny from large surviving trees growing at the same 
location, inoculation of grafts of the large survivors, and inoculation of excised stems of the survivors in 
the laboratory (Griffin et al.l983). Canker length and, in some tests, evaluation of cambium colonization 
and necrosis following canker dissection with a knife, were used to determine disease severit>'. In later 
trials, necrosis at the cambium was evaluated by bark-core sampling of the cankers. To evaluate 
hypovirulence of C. panisilica isolates recovered from the large, surviving trees, the procedure was 
similar except that blight-susceptible, clearcut, or understorv American chestnut trees were inoculated 
with several pathogen isolates from each large, surviving tree. Site factors evaluated included elevation, 
competition from other hardwoods, slope, aspect, and soil characteristics. The results indicated that 
survival of most large trees was associated with low to moderate levels of blight resistance, a low 
frequency of hypovirulence in the blight fungus population on the tree, and favorable sites that were 
relatively free from competition. In 1 5 trials of 13 suniving trees. 73% showed significant e\ idence of 
blight resistance when compared to blight-susceptible reference trees, following //; situ or seedling 
inoculations in the field with standard virulent strains (Griffin et al. 1 983 ). For 3 1 7 blight fungus isolates 
recovered from 19 surviving trees, 28% were hypovirulent or intemiediate hypovirulent in pathogenicitN 
trials on forest American chestnut trees. Several trees exhibited no evidence of blight resistance, and a 
few trees were grow ing in competition w ith other trees. Mites recovered from large survi\ ing trees were 
associated with virulent and h)pov irulent C. panisilica and ma} be agents of hypovirulence spread on 
surviving trees (Wendt et al. 1983: Griffin et al.l984). 



98 



Low disease 
severity 



High disease 
severity 




Environmental factors 

Drought/shallow soil 
High resistance Low winter temperature 

Frost injury 

Hardwood competition/ 
I low light 



High 
susceptibility 



A/TTi Favorable rainfall pattern 
Deep soil 
High light intensity 

Ontogenic factors 
Stem size 
Age 



Figure 1 . Concept of plant disease or chestnut bligiit resistance in relation to observed continuously 
variable disease severity, as represented by the grey scale bar. Low disease severity (whitish top) may be 
called high or complete disease resistance (sensii Vanderplank 1982), high disease severity (blackish 
bottom) called high or complete susceptibility, and intermediate disease severity (medium grey) may be 
called moderate resistance (MR) or partial resistance. For chestnut blight, in any given year or test, 
environmental and ontogenic factors can shift the disease severity rating and resistance rating up or down 
as shown (see arrows). 



BREEDING AMERICAN CHESTNUT FOR BLIGHT RESISTANCE 

Initially, four large, surviving trees (Fig. 2) were selected from this group of surviving trees to start an 
American chestnut blight-resistance breeding program. Controlled intercrosses were made among these 
trees. In this American Chestnut Cooperators" Foundation (ACCF) program, we are testing the 
hypothesis that additive blight resistance can be obtained through two or three cycles of selection. 
Vanderplank (1982) has documented that, in practice, gains of resistance through selection have been 
common and rapid in other host-pathogen systems. The evidence was that this additive resistance is 
mostly oligogenic. In our program, the progeny were grown at high elevation (> 2,500 feet), and after 7 
years, these progeny were evaluated for blight resistance as described above. Two of the crosses, F x M 
and F X G, or the reciprocal cross, produced a high percentage of progeny that had acceptable disease 
severity ratings for further breeding. After one year of canker growth in these crosses, about one-half of 
the progeny trees had relative canker disease severity indexes (canker length x percent necrosis at the 
cambium) that indicated resistance when compared to results obtained for the blight-susceptible clone 
used in the progeny test. Also, this conclusion agrees with data obtained in the /// sifii tests mentioned 
above (Griffin et al. 1983), although strict comparisons to the latter cannot be made. The canker length 
and percent cambium necrosis components of the canker disease severity index for several progeny trees 
and their parents, obtained earlier, are shown in Fig. 2. The M (McDaniel) surviving tree, which had the 



99 



highest canker disease severity index of the four parent trees, combined well v\ith the F (Floyd) surviving 
tree, as did the G (Gault) surviving tree. The WY (Weekly) surviving tree yielded low percentages of 
progeny that had acceptable disease severitv ratings for further breeding. As these trials lasted one year at 
high altitude, they included the winter dormant season, which has been associated with low-tempeiature 
stress, canker development or expansion toward the vascular cambium, and possible breakdown of blight 
resistance at high elevations (see later sections). Using individual F] trees, F2 progeny from the F x M 
and F X G crosses are now being grown. Backcrosses have been made to maintain some of the desirable 
traits of these locally-adapted parent trees. In addition, the progeny of several new intercrosses from trees 
showing high vigor and/or adaptation to different elevations are being gown as additional resistance 
sources. 




W(S) FxM FxM FxM GxF FxG FxG FxWY FxCH FxCH 




LARGE SURVIVOR 
PARENTS 



Figure 2. Canker lengths and percent of cambium areas beneath the cankers that are necrotic (shaded 
portion of bars) in 1 -year duration blight resistance trials, with a standard \irulcnt strain o\W parasitica. 
Tor (below) /// sHii inoculations on branches of the G. I . WY. and M large. sur\ i\ ing American chestnut 
parents, and (above) on main stems for 7-year-old intercross progeny of these surviving trees. The mean 
of seven susceptible reference trees (S) used is shown for the parents and the mean of one W(S) 
susceptible reference tree used for the progeny is shown. CH indicates Chinese chestnut reference parent. 



100 



A large population of American chestnuts with a potential for blight resistance are being grown by ACCF 
cooperators. These trees are progeny from open pollinations of a population of trees derived from large 
survivors and grown in the same breeding orchard. These breeding-orchard trees have exhibited low to 
moderate levels of blight resistance, and many are related to the parent trees used in the control 
pollinations. These trees likely share some of the same alleles that may be responsible for blight 
resistance and other desirable traits. The blight-resistant progeny of these open-pollinated trees should be 
good sources for increasing genetic diversity during later American chestnut restoration efforts. As of 
February, 2004, very large numbers of seedling transplants (93,643) and seed nuts (41,480) have been 
planted, and some have shown promising levels of field resistance to chestnut blight. The best of these 
trees, in terms of durable field blight resistance, can be incorporated into the controlled-cross breeding 
program described above. From these open pollinations, the National Park Service is now growing 4,500 
seedling transplants and 462 trees from nuts. 



INTEGRATION OF BLIGHT RESISTANCE AND HYPOVIRULENCE FOR CHESTNUT BLIGHT 

CONTROL 

In 1980, John Elkins and Bruce Given (West Virginia Department of Agricuhure), of the ACCF, worked 
with Tom Dierauf, of the Virginia Department of Forestry, to graft large, surviving American chestnuts 
on American chestnut rootstocks. These tree rootstocks were established 1 1 years earlier by Al Dietz 
( 1 978), founding officer of the ACCF, at the Lesesne State Forest, for the purpose of breeding a 
blight-resistant American chestnut by radiation breeding. The site in the Lesesne State Forest was 1,350 
feet in elevation and mesic. Most of the trees in this radiation breeding program were dying from 
chestnut blight, but stump sprouts and the rootstocks survived. In 1982 and 1983, blight cankers on the 
American chestnut grafts were inoculated with a mixture of American and European dsRNA-infected, 
hypovirulent strains obtained from J.E. Elliston of the Connecticut Agricultural Experiment Station. Over 
the next two decades, these trees exhibited a high level of blight control, with blight cankers exhibiting a 
high degree of superficiality (Dierauf et al. 1997). By 1999, the largest tree had attained a height of 61 
feet and a dbh of 1 5.7 inches (Griffin 2000). This same tree had a dbh of 1 9.0 inches in March, 2004 
(Fig. 3). This blight control occurred in the presence of an abundant, virulent, ascospore inoculum 
generated by numerous perithecia in the thousands of cankers on the 5,000 American chestnut trees 
planted by Dietz in the Lesesne plantation. Inoculation trials with a standard virulent strain provided 
evidence for blight resistance in these grafts, and the virulent strain was later recovered from the 
superficial cankers resuhing from the inoculations (Robbins and Griffin 1999). Some seedling American 
chestnut trees from the blight-resistant, large survivor intercrosses described above, also have exhibited a 
high to moderate level of blight control after 20 years when inoculated with a European hypovirulent 
strain mixture early or after several years of tree growth. 

Extensive research indicated that dsRNA hypoviruses from the European hypovirulent strains had spread 
into 34-41% of isolates of the blight fungus recovered from the grafted trees at Lesesne, based on colony 
morphologies of over 800 C. parasitica isolates recovered from cankers (Griffin 1999; Robbins and 
Griffin 1999; Hogan and Griffin 2002 a and b). Colonies of European hypovirulent strains commonly 
have a predominantly white phenotype versus the orange-pigmented phenotype of dsRNA-free virulent, 
normal strains or many American hypovirulent strains. Assays of over 70 isolates of C parasitica from 
the grafts, all with a predominantly orange-pigmented phenotype, indicated that most were free of 
hypovirus dsRNA. In white isolates, the European hypovirus, Cryphonectria hypovirus 1 (CHVl, 
Hillman et al. 1995), had spread 642 cm from the hypovirulent-strain-inoculated zone and into a very 
large number (> 45) of vegetative compatibility types of the chestnut blight fungus (Hogan and Griffin 
2002a). Blight resistance may have allowed time for CHVl to spread. Vegetative incompatibility 
between different strains of the blight fungus was believed to be a major barrier to hypovirus transmission 
and biocontrol of chestnut blight in the United States, where a similar large number of vegetative 



compatibility types of C parasitica had been identified in research plots (Anagnostakis and Kranz 1987; 
Anagnostakis and Day 1979; Kiihlman and Bhattyachar>'ya 1984; Lui and Milgroom 1996). Within a 
canker, however, incomplete movement of CHVl within a vegetative compatibilitv tvpe was found for ^ 
the natural cankers on the grafts (Hogan and Griffin 2002b). 

A few predominantly orange-pigmented isolates from the grafts had high dsRNA content, and high 
dsRNA is characteristic of European hypovirulent strains (Dodds 1980). The European hypo\ irulent 
strains inoculated on the grafts were of French and Italian origin. The French strain inoculated on the 




Figure 3. Integrated blight control on American chestnut with resistance and h\po\ irulence. Photo taken 
in March. 2004 of a 19.0-inch dbh stem (left) and an 18.5-inch dbh stem (right) of a two-stem American 
chestnut tree grafted in 1980 from a large survivor and later inoculated with a mixture of hypovirulent 
strains of the blight fungus. Arrow points to a highly superficial (nonkilling) canker. This high level of 
blight control is typical throughout this tree, which is over 60 feet tall. Bars on scale to right arc 1 foot 
long. 



102 



grafts had a predominantly orange-pigmented phenotype, but was derived from a predominantly white 
hypovirulent strain (Elliston 1985). The inoculated Italian hypovirulent strains had predominantly white 
phenotypes. Single-spore analysis of the white isolates from the grafts suggested that hypovirus from the 
Italian inoculated strains had spread on the grafts (Hogan and Griffin 2002a). Colony morphologies of 
hypovirulent strains are variable in subculture, however. Therefore, nucleotide sequence identification 
was used to determine the identity of CHVl in the white isolates and to detennine the identity of CHVl 
in the predominantly pigmented isolates as French or Italian. 

To identify CHVl as French or Italian, hypovirus dsRNA was extracted from predominantly white and 
predominantly pigmented C. parasitica isolates recovered from cankers on the grafts. cDNAs were then 
made by reverse transcriptase-polymerase chain reaction (RT-PCR) for two hypovirus regions: ( 1 ) an 
844-bp region in the helicase domain of open reading frame B (ORF B), and (2) an 894-bp region that 
included part of the 5' non-coding region and part of p29 of ORF A. Nucleotide sequence analysis 
indicated that all pigmented and white C. parasitica hypovirus isolates from the grafted trees and Italian 
inoculated strains had high identities to each other and high identities (98.7-99.9%) to the Italian 
reference hypovirus, CHVl -Euro? (Griffin et al. 2004). Identities to French reference hypovirus 
CHVI-EP713 were low (<89.8%). Thus, no evidence was found for the presence of French hypovirus, 
except for the hypovirus in the predominantly orange-pigmented French hypovirulent strain inoculated on 
the trees. 



FOREST MANAGEMENT IN CLEARCUTS AND PLANATIONS FOR INTEGRATED CHESTNUT 
BLIGHT CONTROL AND RESTORATION OF AMERICAN CHESTNUT 

American chestnut presently may be found sparsely to frequently as an understory tree throughout the 
native range of the species in both the former oak-chestnut and mixed mesophytic forest regions (Braun 
1950). Xeric and intermediate sites may have very high population densities of understory American 
chestnuts, especially at elevations of about 3,000 feet or higher in the Mid-Atlantic area of the fonner 
oak-chestnut region. Highly mesic sites may have little or no understory American chestnut survival 
(Griffin 1992a). Chestnut blight is endemic in these understory trees with about 15-20% blight incidence. 
When these areas are clearcut, American chestnut grows as rapidly as any hardwood in the clearcut 
(Smith 1977). This rapid growth is followed by a chestnut blight epidemic over a 10-year period when 
90-100% of the chestnut trees are blighted. A great abundance of virulent ascospore inoculum develops 
from perithecia on hundreds of American chestnut stems in these clearcuts. Similar epidemics occur in 
blight-susceptible American chestnut plantations. Near the end of the epidemic, some trees exhibit 
superficial cankers which are associated with hypovirulent strains, some possibly originating from 
cankers in the understory American chestnuts (Griffin et al. 1983; Griffin et al. 1984). On xeric and 
intermediate sites, following stem death from blight, numerous stump sprouts develop, some of which are 
browsed by deer. On mesic sites, with high competition from hardwoods, few stump sprouts develop, 
almost all of which are browsed by deer. On these sites with great tree-growth potential, American 
chestnut rootstocks are completely lost. Survival of American chestnut over all sites was inversely related 
to basal area of competing hardwoods (Griffin et al. 1991 ). Light intensities were very low on sprouts at 
the base of the stump on the mesic sites (Griffin 1992b). 

Forest management involving removal of competing hardwoods resulted in the development of superficial 
cankers that were associated with dsRNA-infected hypovirulent strains of the chestnut blight fungus 
(Griffin et al. 1991; Griffin et al. 1993). This management practice increased stem size, promoted mast 
production, and maintained the survival of chestnut stems for several years beyond that found in check 
plots having no removal of competing hardwoods. The greatest blight control was found on a mesic site 
that was clearcut and a mesic plantation site (Griffin et al. 1991 ). However, at all locations blight control 



103 



eventually broke down. Blight control also broke down in clearcuts where hypovirulent strains were 
artificially introduced. Research indicated that this breakdown of biocontrol was associated with the 
following: 1) the high blight susceptibility (quick kill) of forest American chestnuts; 2a) the secondary 
colonization of superficial (hypovirulent) cankers b\ virulent strains in dixerse vegetative compaiibilit\ 
types that were generated in the clearcul or plantation; 2b) the development of new killing cankers 
elsewhere on the chestnut stem by virulent strains; 3) the slow spread of hypovirulence; and 4) breakdown 
at high altitude of superficial (hypovirulent) cankers over winter (Griffin et al 1993; Griffin and Griffin 
1995). In the absence of hardwood management, even in unmanaged plantations, factor 5) is high 
hardwood competition and the associated reduced light. All lead or contribute to chestnut stem death. 
Using artificially introduced hypovirulent strains in forest situations, others have also found blight control 
to be either unsuccessful (Liu et al. 2002) or partial (Anagnostakis 2001 ). Further, introduced European 
hypoviruses (CHVl) did not persist at sites where they were introduced (Liu et al. 2002; Peever et al. 
1997). 

In our study, the clearcut and plantation blight-susceptible trees grew at altitudes ranging from 2,000 to 
3,500 feet, hi contrast, some blight-susceptible American chestnuts naturally infected with hypovirulent 
strains at low altitudes (< 1,000 feet elevation) have exhibited durable blight control, even in the absence 
of blight resistance (Griffin et al. 1983; Griffin and Griffin 1995). When pure cultures of hypovirulent 
strains and bark plugs from a low altitude tree were inoculated into American chestnut trees at high 
altitude (3,500 feet in elevation), superficial cankers were produced during the growing season. However, 
the hypovirulent strains colonized the cambium over the winter, causing a breakdown in the canker 
superficiality rating and biocontrol (Griffin and Griffin 1995). Tests indicated that the hypovirus in the 
hypovirulent strains survived the winter. These trees died as the cambium was completeK killed. 
Conversely, the original low-altitude American chestnut tree (hypovirulence source) still exhibited stable 
blight control as of March, 2004 (unpublished). 

As indicated above, physiological stress in chestnut species may occur at high altitudes, from early frosts, 
and in frost pockets. At these sites, blight severity can be very high even on the highly blight-resistant 
Chinese chestnut and some stems can be killed (Berry 1951; Headland et al 1976; Jones et al. 1980). 
Other studies have indicated that American chestnut at high altitude sites are under ph\ siological stress. 
This is indicated by greatly increased electrolyte leakage from bark tissues collected from high altitude 
(3,900 feet) versus low altitude (530 feet) sites (Griffin 2000). The above findings suggest that the blight 
control breakdown at high-altitude clearcuts described above may be related in part to physiological 
stress. Some large-surviving American chestnut trees have been found at elevations higher than 4.000 
feet, and these trees, adapted to blight at these elevations, may be useful for breeding and integrated blight 
control at higher elevations. High elevation sites (>2,500 feet elevation) ma\ account for the bulk of the 
surviving population of understor> American chestnuts in Virginia 

Physiological stress on American chestnuts may also occur during the summer drought periods commonly 
encountered in the former oak-chestnut forest region (Braun 1950). Gao and Shain (1995) found drought 
stress may contribute to blight susceptibility. Additionally Anagnostakis (2001 ) found that in 
Connecticut forest plots, where hypovirulence was ailificiall\ introduced. American chestnut trees with 
many cankers often died in the summer following drought the pre\ ious year. As indicated abo\e. 
biocontrol associated with natural hypovirulence was less on managed, xeric clearcut sites than on 
managed, mesic clearcut sites (GrifTm et al. 1991 ). Xeric slope sites frequently have small American 
chestnut stumps and very high population densities of understor> American chestnuts that are associated 
with a less dense canopy. This can lead to the false conclusion that the\ are the best chestnut restoration 
sites. In contrast, mesic shallow cove sites on slopes had large American chestnut stumps (up to 3-4 feet 
in diameter) with moderate numbers of understor}. American chestnuts ((iriffin 1992a). Deep cove sites 
had medium-sized stumps and little or no understor> American chestnuts. Mesic shallow cove sites on 
slopes have great potential for chestnut growth, integrated blight control w ith resistance, hypovirulence. 



104 



and forest management, chestnut restoration, and natural regeneration of American chestnut. Often they 
have deep, fertile soils. Large-surviving American chestnuts at high altitude (>2,500 feet) may be useful 
for breeding and integrated control at higher elevation, mesic sites. 



CONCLUSIONS 

Research indicated that the survival of most large surviving American chestnut trees was associated with 
resistance, hypovirulence, and favorable sites. Our ACCF breeding program utilizes controlled 
intercrosses of these trees, which has resulted in progeny with acceptable levels of resistance. These 
progeny trees have been used in further breeding. This may result in trees with additive blight resistance. 
Some American chestnuts in a very large population (135,123 trees), derived from open pollinations of 
large survivors, have shown promising levels of field blight resistance. This large population may serve 
as a source of genetic diversity in future restoration efforts. The high blight susceptibility of forest 
American chestnuts, along with the abundance of a vc-diverse, virulent inoculum of C parasitica has 
severely limited blight control and the use of hypovirulence in forest clearcuts and plantations. High 
altitude, low temperature stress, drought stress, and hardwood competition are additional factors that 
inhibit blight control and the use of hypovirulence. However, a long term (> 20 years) and high level of 
blight control has been obtained on mesic, hardwood-managed sites. These sites were established with 
blight-resistant American chestnut trees that were inoculated with hypovirulent strains of C. parasitica. 
Nucleotide sequence analysis indicated that blight control was associated with the spread of Italian 
CHVl . Some blight resistance may be needed in American chestnut to allow time for hypoviruses to 
spread. Site selection and removal of competing hardwoods may be critical forest management practices 
needed for blight control. Mesic shallow coves on lower altitude slopes are among the best sites to 
implement integrated use of resistance, hypovirulence, and forest management for blight control and 
restoration of American chestnut. High-altitude (>2,500 feet), large-surviving American chestnuts may 
be useful for breeding and integrated control at higher elevation, mesic sites. 



ACKNOWLEDGEMENTS 
We thank E. Hogan and J. Eisenback for assistance with graphics. 

LITERATURE CITED 

Anagnostakis, S.L. 2001 . American chestnut sprout survival with biological control of the chestnut-blight 
fungus population. For. Ecol. and Man. 15:225-233. 

Anagnostakis, S.L. 1992. Measuring resistance of chestnut trees to chestnut blight. Can. J. For. Res. 

22:568-571. 

Anagnostakis, S.L., and P.R. Day. 1979. Hypovirulence conversion in Endothia parasitica. 
Phytopathology 69: 1 226- 1 229. 

Anagnostakis, S.L., and J. Kranz. 1987. Population dynamics of Cryphonectria parasitica in a mixed- 
hardwood forest in Connecticut. Phytopathology 77:751-754. 

Bazzigher, G., and P. Schmid. 1962. Metodik zur Prufung der £A?(io//7/fl-Resistenz bei Kastanien. 
Phytopathol.Z. 45:169-189. 



105 



Braun, L. 1950. Deciduous forests of eastern North America. Blackiston. Philadelphia, PA. 595 p. 

Berpy. F.H. 195 1 . Winter iniur> to Asiatic chestnut trees in the South during November. 1950. Plant Dis. 
Rep. 35:504-505. 

Clapper, R.B. 1952. Relative blight resistance of some chestnut species and hvbrids. J. For. 50:453-455. 

Dierauf T., J. Artman. J.R. Elkins, S.L. Griffin, and G.J. Griffin. 1997. High level of chestnut blight 
control on grafted American chestnut trees inoculated with hypovirulent strains. J. Arbor. 23:87-88. 

Dietz, A. 1978. The use of ionizing radiation to develop a blight resistant American chestnut, Castanea 
denkita, through induced mutations. P. 17-20 in Proc. of the American chestnut symposium, McDonald, 
W.L. et al. (eds.). West Virginia Books, Morgantovvn, WV. 

Dodds, J. A. 1980. Revised estimates of molecular weights of dsRNA segments in hvpovirulent strains of 
Endothia parasitica. Phytopathology 970: 1 2 1 7-1220. 

Elliston. J. 1985. Characteristics of dsRNA-free and dsRNA-containing strains of Etidothia parasitica in 
relation to hypovirulence. Phytopathology 74:151-158. 

Goa, S., and L. Shain. 1995. Effects of water stress on chestnut blight. Can. J. For. Res. 25:1030-1035. 

Griffin. G.J. 2000. Blight control and restoration of the American chestnut. J. For. 98:22-27. 

Griffm, G.J. 1992a. American chestnut survival in understory mesic sites following the chestnut blight 
pandemic. Can. J. Bot. 70:1950-1956. 

Griffin. G.J. 1992b. Irradiance characteristics associated with American chestnut survival in forest 
clearcuts and understory mesic sites. Revista Academia Galega de Ciencias (Spain) 1 1 :5-14. 

Griffin, G.J.. and J.R. Elkins. 1986. Chestnut blight. P. 1-26 in Chestnut blight, other Endothia diseases, 
and the genus Endothia. APS Press, St. Paul, MN. 

Griffin. G.J. and S.L. Griffin. 1995. Evaluation of superficial canker instabilitv for hypovirulent 
Cryphonectria parasitica inoculated on American chestnut trees. Eur. J. For. Path. 25:351-355. 

Griffin. G.J., F.V. Hebard. R.W. Wendt. and J.R. Elkins. 1983. Survival of American chestnut trees: 
evaluation of blight resistance and virulence in Endothia parasitica. Phytopathologv 73:1084-1092. 

Griffin, G.J.. M.A. Kahn, and S.L. Griffin. 1993. Superficial canker instabilitv during winter and 
virulence of Endothia parasitica associated with managed forest clearcut and plantation American 
chestnut trees. Can. J. Plant Pathol. 15:159-167. 

(jriffin G.J., N. Robbins, E.P. Hogan. and G. Farias-Santopietro. 2004. Nucleotide sequence identification 
of Cryphonectria hypovirus 1 infecting Cryphonectria parasitica on grafted American chestnut trees 12- 
18 years after inoculation with a hypovirulent strain mixture. For. Path. 34:33-46. 

Cjriffin. Ci.J.. 1 l.C. Smith, A. Dietz. and J.R. lilkins. 1991 . Importance of hardwood competition to 
American chestnut surv ival, growth, and blight development in forest clearcuts. Can. J. Bot. 
69:1804-1809. 



106 



Griffin, G.J., R.A. Wendt. and J.R. Elkins. 1984. Association of hypovirulent Enduthia parasitica with 
American chestnut in forest clearcuts and with mites. (Abstr.) Phytopathology 74:804. 

Headland, J.K., G.J. Griffin, R.J. Stipes, and J.R. Elkins. 1976. Severity of natural Endothia parasilica 
infection of Chinese chestnut. Plant Dis. Rep. 60:426-429. 

Hebard, F.V. 1999. Meadowview notes 1998-1999. J. Am. Chestnut Found. 13:7-15. 

Hillman, B.I., D.W. Fulbright, D.L. Nuss, and N.K. Vanalfen. 1995. Hypoviriddia. P. 261-264 in Report 
Int. Committee Taxon, Viruses, 6"' edn. Murphy, F.A. et al. (eds.). Springer-Verlag. 

Hogan, E.P., and G.J. Griffin. 2002a. Spread of Ciyphonecfria hypovirus I into 45 vegetative 
compatibility types oi Cryphonectria parasitica on grafted American chestnut trees. For. Path. 32:73-85. 

Hogan, E.P., and G.J. Griffin. 2002b. Incomplete movement of Cryphonectria hypovirus 1 within a 
vegetative compatibility type of Cryphonectria parasitica in natural cankers on grafted American 
chestnut trees. For. Path. 32:331-344. 

Kuhlman, E.G., and H. Bhattacharyya. 1984. Vegetative compatibility and hypovirulence conversion 
among naturally occurring isolates of Cryphonectria parasitica. Phytopathology 74:659-664. 

Jones, C, G.J. Griffin, and J.R. Elkins. 1980. Association of climatic stress with blight on Chinese 
chestnut in the eastern United States. Plant Dis. 64:1001-1004. 

Liu, Y.-C, M.L. Double, W.L. MacDonald, and M.G. Milgroom. 2002. Persistence of Cryphonectria 
hypoviruses after their release for biological control of chestnut blight in West Virginia forests. For. Path. 
32:345-356. 

Liu, Y.-C, and M.G. Milgroom. 1996. Correlation between hypovirulence transmission and the number 
of vegetative incompatible (vie) genes different among isolates from a natural population of 
Cryphonectria parasitica. Phytopathology 86: 79-86. 

Peever, T.L., Y.-C. Liu, and M.G. Milgroom. 1997. Diversity of hypoviruses and other double-stranded 
RNAs in Cryphonectria parasitica in North America. Phytopathology 87:1026-1033. 

Robbins, N., and G.J. Griffin. 1999. Spread of white hypovirulent strains of Cryphonectria parasitica on 
grafted American chestnut trees exhibiting a high level of blight control. Eur. J. For. Path. 29:51-64. 

Smith, H.C. 1977. Height of tallest saplings in 10-year-old Appalachian hardwood clearcuts. USDA For. 
Serv. Res. Pap. NE-381. 

Uchida. K. 1977. Studies on Endothia canker of Japanese chestnut trees caused by Endothia parasitica 
(Murrill) P.J. ed. H.W. Anderson. Bull. Ibaraki-Ken Hortic. Exp. Stn. Spec. Issue 4 (Japan). 65 p. 

Vanderplank, J.E. 1982. Host-pathogen interactions in plant disease. Academic Press, New York. 207 p. 

Wendt, R., J. Weidhaas, G.J. Griffin, and J.R. Elkins. 1983. Association of Endothia parasitica with 
mites isolated from cankers on American chestnut trees. Plant Dis. 67:757-758. 



107 



108 



Steiner, K. C. and Carlson, J. E, eds. 2006. Restoration of American Chestnut To Forest Lands - Proceedings of a 
Conference and Workshop. May 4-6, 2004, The North Carolina Arboretum. Natural Resources Report 
NPS/NCR/CUE/NRR - 2006/001, National Park Service. Washington, DC. 

GENETIC STRUCTURE OF AMERICAN CHESTNUT POPULATIONS BASED ON 

NEUTRAL DNA MARKERS 

"^ Thomas L. Kubisiak and James H. Roberds 

USDA Forest Service. Southern Research Station, Southern Institute of Forest Genetics, 
23332 Hwy 67, Saucier, MS 39574 USA (tkubisiak@fs.fed.us) 



Abstract: Microsatellite and RAPD markers suggest that American chestnut exists as a highly variable 
species, even at the margins of its natural range, with a large proportion of its genetic variability occurring 
within populations (-95%). A statistically significant proportion also exists among populations. 
Although genetic differentiation among populations has taken place, no disjunct regional pattern of 
variation exists. A cline in allele frequencies and number of rare alleles occurs along the Appalachian 
axis, with the highest levels of gene diversity and the greatest numbers of rare alleles being found in 
southwestern populations. Population pairwise estimates of genetic distance are significantly associated 
with the geographic distance between populations. Geographically proximate populations are slightly 
more genetically similar than geographically distant populations. Genetic variability in American 
chestnut follows a pattern consistent with the hypothesis of a single metapopulation in which genetic drift 
plays a major evolutionary role. Results of this study are based on neutral genetic loci and do not 
necessarily reflect genetic differentiation for adaptive genes or gene complexes. Therefore, in order to 
assure that most of the variation at these genes is also captured in conservation and breeding endeavors, 
sampling should focus on collecting a fairly large number of individuals from each of several geographic 
areas. 

Keywords: Castanea dentata I SSR / RAPD / genotypic diversity / haplotype diversity 



INTRODUCTION 

The American chestnut {Castanea dentata Borkh.) was once one of the most important timber and nut- 
producing tree species in eastern North America (U.S. Census Bureau 1908). Its native range extended 
from southern Maine and Ontario in the north to Georgia, Alabama and Mississippi in the south (Sargent 
1905). The species now exists primarily as stump sprouts across this entire range, the victim of a 
devastating canker disease. The disease, chestnut blight, is caused by an exotic fungal pathogen now 
known systematically as Cryphonectria parasitica (Barr 1979). After more than half a century of blight, 
numerous living stems of American chestnut still exist in the understory of upland forests in the mid- 
Appalachians (Stephenson et al. 1991 ). Prolific stump sprouting has enabled American chestnut to 
persist, but as sexual reproduction is infrequent, its gene pool will likely face serious erosion when old 
root systems fail to produce sprouts and perish. 

Because resistance to C parasitica is low or lacking in American chestnut, Burnham (1981 ) proposed the 
use of a classical backcross breeding program to develop blight resistant timber-type trees. Adopting this 
methodology as their charter, the non-profit philanthropic organization The American Chestnut 
Foundation (TACF) has since developed a vigorous backcross breeding program designed to introgress 
the resistance of Chinese chestnut (C. moUissima Blume) into American chestnut (Hebard 1994; Kubisiak 
et al. 1997). TACF's initial efforts focused on American chestnut trees in southwest Virginia, but the goal 
is to restore the species throughout its entire native range. Thus, information regarding the amount and 
distribution of molecular genetic variation in American chestnut might help to better determine the 
number of breeding locations that will be needed across the species range. 



109 



Previously, little was known about how genetic variability is distributed across the landscape that - ^ 
comprises the natural range of this species. In an exploratorv examination of genetic variability for ~ 

American chestnut, Huang et al. ( 1998) obtained results with alloz>me and random amplified 
polymorphic DNA (RAPD) markers that suggest as many as four regional metapopulations might exist. 
However, hierarchical AMOVA was not performed to quantify this putative regional component, nor 
were statistical tests employed to test for significant differences. Since that research was completed, the 
magnitude, significance, and patterns of regional structure have been the subjects of much discussion and 
debate (F.V. Hebard, P. Sisco, and G. Miller personal communication). Given the importance of regional 
structure in regards to breeding blight resistant regionally adapted American chestnut, we felt compelled 
to embark on a more thorough examination of genetic variation in American chestnut using microsatellite 
and RAPD markers. 

Here we report results obtained from an analysis of genetic structure for populations of American 
chestnut occurring over a significant portion of its natural range. We assa>ed six microsatellite and 19 
RAPD markers and based our analysis on allele and haplotype frequency variation observed for these 
neutral loci. Our objective for this research was to secure a more detailed and complete understanding of 
population structure for American chestnut. In the following sections we describe genetic differentiation 
patterns observed within and among populations and report estimates of diversitv parameters associated 
with microsatellite and RAPD loci segregating in American chestnut. Finally, we compare our results to 
patterns of variability previously reported for neutral markers in American chestnut as well as in other 
tree species. 



MATERIALS AND METHODS 
Population sampling and DNA extraction 

A rangewide sampling of expanded leaves or dormant buds of American chestnut were collected at 22 
sites across its natural range (refer to Figure I ). Most of the samples were collected from sites in State or 
National Forests, but a few sites were located on private land holdings. Each sample was assigned a 
unique ID and sent to the Southern Institute of Forest Genetics in Saucier. Mississippi for DNA extraction 
and analysis. Total nucleic acids were isolated from tree tissues as described in Kubisiak et al. (1997). 

Species evaluation 

A panel of DNAs consisting of eight American chestnut {one from each of eight different sites sampled 
tor this study), six Chinese chestnut (trees from USDA import #'s 70315, 10406 1. 78626, 104014, 
104015, and 104016), seven Henry chinkapin (C. henryi Rehder & Wils.) (trees from USDA import U 
104058, the Nanjing Botanical Garden, Nanjing, Peoples Republic of 

China (PRC), and the Wuhan Institute of Botany, Wuhan, PRC), four Seguin chestnut (C. seguinii Dode) 
(trees from USDA import # 703 1 7). seven European chestnut (C. saliva Mill.) (including trees from the 
Caucasus Mountains of southern Russia. Bursa, lurkey, and the Black Forest in German) ). and eight 
Alleghan> chinkapin (C. piimila Mill.) (Harrison County. Mississippi) were amplified using the 
polymerase chain reaction (PCR) and a chloroplast-specific primer pair (a, b) as described in Taberlet et 
al.(199l). 




Figure 1 . Map of the geographic origin of the 22 Castanea dentata Borkh. populations sampled in this investigation. 
The number in parentheses refers to the number of trees sampled at each location. 



Microsatellite PCR amplification and detection 

Primer sequences and PCR conditions for microsatellite loci developed in European chestnut (C. sativa) 
were obtained from the literature (Marinoni et al. 2003). Primer sequences for microsatellite loci 
developed in white oak {Quercus alba L.) were obtained from A. David and D. Wagner at the University 
of Kentucky. For each microsatellite, the forward primer was 5'-end labeled with one of three f.uorescent 
dyes to facilitate detection using the Applied Biosystems 3 1 00 Genetic Analyzer and the GEN ESC AN" 
version 3.7 fragment analysis software (Applied Biosystems, Inc. Foster City, CA). Microsatellites were 
PCR amplified and the products post-PCR multiplexed by color and size whenever possible. Allele sizes 
were determined by including the GENESCAN*-500[TAMRA] internal size standard in each sample 
lane. The data were scored using GENOTYPER" version 3.7 (Applied Biosystems, Inc. Foster City, 
CA). 



RAPD PCR amplification and detection 

RAPD amplification and detection was based on the protocols reported in Kubisiak et al. (1997). RAPD 
fragments were identified by the manufacturer primer code corresponding to the primer responsible for 
their amplification, followed by a subscript four digit number indicating the approximate fragment size in 
base pairs. Markers were chosen based on the intensity of amplification (only intensely amplified bands 
were scored) and the absence of co-migrating DNA fragments. All markers were found to conform to 
Mendelian expectations based on their inheritance in at least one of four different interspecific chestnut 
pedigrees. 



Data analysis 

A search for common/redundant multi-locus genotypes and haplotypes was performed (Excoffier and 
Slatkin 1995). Populations were tested for Hardy- Weinberg proportions using both % and G" tests (Weir 
1990). Allele frequencies for each population were computed and estimates obtained for effective 
number of alleles per locus (Ae), Nei's (1972) measures of gene diversity (h), Nei's (1978) unbiased 



11 



measure of genetic distance (D). Michalakis and Excoffier's (1996) genetic differentiation measure (Ost) 
for the microsatellite loci, and Nei's (1987) genetic differentiation measure (Gst) for RAPD loci using the 
software program ARLEQUIN version 2.001 (Schneider et al. 2000) and POPGENE version 1.31 (Yeh et 
al. 1997). In addition, x" and G^ tests were calculated to test homogeneity of allele frequencies among 
populations. For microsatellite analysis in ARLEQUIN, alleles were coded assuming a step-wise 
mutation model. Associations between allele frequency and latitude or longitude were first studied using 
the PRCX STEPWISE procedure in SAS version 8.0I(SAS, 1999). A variable was only added to the 
model if its F-statistic was significant at the 5% level. Once added, any variable that did not have a F- 
value significant at the 5% level was deleted from the model. Associations between the observed number 
of alleles per locus, number of rare alleles per locus (rare alleles are those with frequencies less than 0.05 
computed across all populations), effective number of alleles per locus, gene diversity and latitude or 
longitude were also studied. In order to further investigate any apparent clinal trends, a composite 
dependent variable (CDV) was computed that combined both latitude and longitude. First, a reference 
line was drawn between the southwestern most and northeastern most populations. Then, perpendicular 
lines were drawn that connected the various populations to this line. Distances (converted into 
kilometers) along the reference line to the population perpendiculars were used as values for CDV. 
Genetic distance (D) and among population differentiation were calculated for each pair of populations 
and associations with geographic distance were investigated using the PROC REG procedure in SAS. 
Genetic associations existing among populations were first studied using unweighted pair-group mean 
analysis (UPGMA) based on the matrix of Nei's genetic distance, and then by principal components 
analysis (PCA) conducted on allele frequency data using the PROC PRJNCOMP procedure in SAS. 



RESULTS 
Putative species identification 

Primers that amplified the intergenic spacer region between tniT (UGU) and the tniL (UAA) 5' exon of 
the chloroplast genome (primers a and b: 5'-CATTACAAATGCGATGCTCT-3' and 5'- 
TCTACCGATTTCGCCATATC-3', respectively: Taberlet et al. 1991 ) were found to uniquely 
differentiate American chestnut chloroplast DNA from all other Castaiiea (chestnut and chinkapin) 
species. Based on DNA sequence data (data courtesv F. Dane and P. Lang of Auburn Universitv ) this 
primer pair was found to amplify a band 857 base pairs (bp) in length in American chestnut, and bands 
ranging from 942 to 945 bp in all other Castauea species including the native chinkapin (both C. pumila 
var. alleghaniensis and C. pumila var. ozarkeiisis). Much of the size difference observed between 
American chestnut and the other Castauea species was due to two unique deletions (one 12 bp and the 
other 75 bp in length) contained within this region of the American chestnut chloroplast genome. A 
larger sampling of native chinkapin (specifically ('. punii/a: var. alle^hauicusis - 48 trees) has \et to show 
the presence of these large deletions. 

Based on the phenotype observed for this marker, of the 1 158 chestnut trees sampled for this studv 165 
(14.2%) were eliminated from further analysis as they did not hav e the smaller chloroplast band 
characteristic of American chestnut. These 165 trees were collected from nine different sample sites. 
Four of the nine sites had very few suspect trees. One site had to be completeh eliminated from the stud} 
as all trees sampled were found to be suspect. Four sites had to be pooled with the most geographically 
proximate site in the same state as a large number of suspect trees were found. In total, as many as 993 
trees from 1 8 different sample sites were available for analysis of genetic variation in American chestnut. 



Microsatellite-based genetic differentiation 

Data describing the microsateliite loci used in our analyses are presented in Table 1 . Considerable 
variation was displayed by the 6 loci. Five of the six loci had very little missing data and were thus used 
to search for common or redundant multilocus genotypes (MLGs) and haplotypes. Based on these five 
loci, only five redundant MLGs were observed. Each redundant MLG was only found to occur twice. 
Based on the same five loci, 1 14 of 1603 estimated haplotypes were found to occur more than once either 
within or across populations, but there was no apparent geographic trend to their distribution. 



Table 1 . Microsateliite and RAPD primer sequence, repeat type, allele size, and number 
of unique alleles identified in samples collected from 18 populations oi Castanea dentata 
Borkh. located throughout the species natural range in eastern North America. 



Locus 



Primer 
Sequence 5'-3' 



Repeat 
type 



Allele size 
(bp) 



Number of 
unique alleles 



Microsatellites 
CiC ATOP 

C5CATI4 

C5CATI5 

QaCAOll 

OaGA068 

C'aGA209 

RAPDs 
106 



F": AGAATGCCC ACTI TTGCA 

RCTCCCTTATGGTCTCG 

FGAGGTTGTTGTTCATCATTAC 

R:ATCTCAAGTCAAAAGGTGTC 

F : TCTGCG ACCTCG AA ACCG A 

RCTAGGGTTTTCATTTCTAG 

F:AACAATAGGAGTTGGTTTGAG 

RGTTAGGGTTTGGAAAATAGGA 

FGCTTTTCTTTCCAGGGCTAC 

R:GTGGGACAGTGAGGCAGAG 

F : C AAGC AGTATTGTTTTATCTC 

RGTTGCCCCTGTGAACTAC 

CGTCTGCCCG 



184 


CAAACGGCAC 


213 


CAGCGAACTA 


225 


CGACTCACAG 


237 


CGACCAGAGC 


423 


GGGTCTCGAA 


500 


TTGCGTCATG 


514 


CGGTTAGACG 



(AC),AT(AC)„ 


167-211 


(AC)„ 


121-151 


(AG)n 


115-141 


(AC)n 


160-188 


(AG)„ 


156-192 


(AG)„ 


227-265 


NA 


500 




525 




650 




700 




800 


NA 


450 




1150 




1800 


NA 


900 


NA 


1000 


NA 


800 




1450 


NA 


825 




1000 




1250 


NA 


600 




875 


NA 


775 


NA 


575 



31 
15 
15 
13 
17 
15 



^Locus names beginning with Cs were derived from Castanea sativa (Marnioni et al 2003) and those beginning with Qa 
were derived from Oiiercus alba (sequences courtesy of A David and D Wagner). RAPD primer sequences were obtained 
from J Hobbs at the University of British Columbia. BC, Canada 
''F=forvvard primer, and R=reverse primer 

The expected genotype frequencies at all loci, and in all populations, conformed to Hardy- Weinberg 
expectations, except for locus ^aGA209 in population PCKY that showed a significant excess of 
homozygotes. Frequencies for alleles at greater than 1 0% frequency over all populations, plus those 
found to be significantly associated with latitude and/or longitude, are displayed by population in Table 2. 



113 



All six single-locus contingency x" analyses as well as G" tests for homogeneitv of allele frequency across 
populations indicated significant (p<().05) departures from homogeneity'. 



Table 2. Microsatellite allele frequencies estimated for 1 8 populations of Castanea dentata 
Borkli. located throughout the species natural range in eastern North America. Only those 
alleles at frequencies greater than 0. 1 o\er all populations and those alleles significantly 
associated with latitude or longitude (identified in italic and bold-italic, respectively) are 
presented. 



Locus and 










,\C^! 
















<^^ 










Allele (bp) 


CCNC 


BCNC 


GCSC 


PCKV 


RCkY 


SGCVA 


ONTCA 


PCWV 


GCMD 


WCPA 


VCHA 


MPCPA 


UCN\ 


LCCT 


RCNY 


HCMA 


MCCT 


ME 


CcCATOl 






































ik: 


() 030! 


0imn 


0517 


0096 


mm 


0.0424 


0.0000 


00067 


00000 


0000 


0000 


0000 


00000 


035 7 


OttOtJ/t 


0179 


0200 


00000 


;«6 


0MS5 


O0O9S 


0.0172 


0.10511 


0.1154 


0.0508 


0.0172 


0.0467 


0.0000 


0.01-9 


0.0000 


0.0294 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


IS? 


0.0000 


0.0196 


00000 


0.0000 


O.tHlOO 


O03.t9 


0.0603 


0.026- 


0.0693 


0.0714 


0.035- 


0.0980 


0.1489 


0.0536 


0.1379 


0.2232 


0.0900 


0.1125 


190 


OIU^S 


II 17(,5 


II inu 


110^77 


iiiri.'^ 


(1 11591 


OIKVXI 


OlliW.7 


1) isi: 


(nri4 


0(1804 


II i:-5 


(1 rii: 


0:141 


3448 


0(1357 


IINHI 


15(1(1 


m 


0.0000 


0.0000 


0.05 1 7 


0.0096 


0.0000 


0.0000 


0.0603 


0.0133 


0.O990 


0.0089 


0.01-9 


0.0196 


0.095- 


01429 


0.1034 


0.0446 


0.0500 


0.0625 


192 


01439 


01569 


0.01162 


0.1346 


0.0'69 


0.0593 


0.0517 


0.0600 


0.0545 


0.0089 


0.0089 


0.0000 


0.1064 


0.0089 


0.0690 


0.0089 


0.0000 


0.0125 


194 


0985 


1275 


(Ki90 


0769 


o:ii5 


0761 


1638 


1467 


0(.93 


I607 


1016 


0980 


0957 


2679 


00776 


0<.25 


0700 


0875 


196 


1409 


1173 


2241 


2692 


1518 


1864 


1897 


1911 


1218 


11 3929 


1929 


2647 


2872 


02411 


I46(. 


1429 


0X00 


IJOO 


200 


0.0612 


0.05M 


0.0345 


0.04SI 


0.0962 


0.1102 


0.0862 


0.0400 


0.0050 


0.0000 


0.0000 


0.0490 


0.0000 


0.0089 


0.0259 


0.0089 


0.0000 


0.03-5 


202 


OOO^r, 


oovs 


0690 


0000 


omut 


nooti 


fXiOO 


I)ft6 " 


IHl5tt 


fHino 


026S 


0000 


01X10 


OOf to 


00000 


00000 


00000 


00000 


CsCATM 






































129 


1(145 


1371 


1552 


01458 


0.1077 


0I7S0 


0250U 


024O3 


02892 


02632 


0366I 


1389 


3043 


1455 


02500 


1792 


04314 


1750 


\n 


3284 


W78 


4481 


4792 


4808 


5000 


043 10 


(14091 


4069 


3 1 58 


01661 


3796 


2147 


6IHK) 


05690 


04717 


3725 


05125 


14} 


0000 


0000 


0000 


ixmi 


OOilfl 


0000 


0259 


Of too 


tun 96 


tUiOO 


0001) 


00000 


021- 


00000 


OOOSfi 


0566 


0000 


00125 


US 


0.007$ 


0.009S 


0.0000 


0.0104 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


CsCATl? 






































11^ 


0104 


0IH5 


0000 


00000 


Ofmoo 


ooono 


00000 


0000 


00000 


0000 


0000 


0000 


0000 


00(10 


00000 


00000 


0000 


0000 


IP 


0104 


0926 


1250 


00135 


;«< 


0X33 


00000 


01346 


0309 


0122 


0167 


0132 


00000 


0000 


00000 


00000 


0000 


0357 


121 


00104 


nil 


0750 


1622 


I53X 


OH33 


01129 


0.M.5 


203- 


2H05 


O0333 


01316 


22-3 


I3S3 


3021 


01410 


0500 


01 -S6 


127 


0521 


0556 


0500 


0270 


0" 69 


O.0S33 


00000 


03^5 


0000 


0000 


0000 


Oftltfl 


«"<« 


0000 


00000 


012'i 


00000 


0000 


133 


2708 


1111 


1250 


1486 


1154 


KMNI 


0.2097 


0.254HI 


1852 


1951 


0.1333 


02168 


01818 


1064 


1(M2 


2436 


021X10 


02500 


137 


0521 


I4HI 


0000 


0405 


0.^s5 


05011 


019S5 


12.10 


2037 


ns54 


2000 


01316 


1212 


02766 


03646 


3333 


1000 


02500 


139 


0.1042 


0.0370 


0.0000 


0.1 OSI 


0.0-69 


0.1000 


0.0806 


0.1154 


0.0370 


0.0854 


e.eis«7 


0.1184 


0.0152 


0.0319 


0.0000 


0.0128 


0.0125 


0.0357 


QaCA022 






































170 


4141 


6058 


a 4655 


2885 


251HI 


2500 


2586 


0:1:1 


5727 


05603 


03482 


n 3627 


468S 


4286 


0474 1 


2411 


2000 


2683 


172 


0.1797 


0.230t 


0.2414 


0.2596 


OI53S 


0.2679 


0.1810 


0.2808 


0.1 591 


0.1207 


0.1964 


0J824 


0.3125 


0.3661 


0J362 


0JO36 


0.5100 


0.3049 


174 


0312 


0192 


0345 


0192 


OO.i.'<5 


0446 


ooaso 


00137 


a 00011 


00H6 


O0OH9 


0000 


00000 


00000 


00000 


00000 


0000 


00122 


m2 


0234 


028K 


0172 


00759 


0962 


2232 


02672 


1438 


0955 


0517 


1786 


0686 


01)312 


0.1071 


00862 


00893 


OKKX) 


OI461 


QaGA068 






































164 


075H 


0000 


1379 


0943 


0400 


0000 


00000 


0065 


0140 


0000 


00000 


00000 


oono 


00000 


00000 


0000 


0000 


0122 


166 


2955 


1250 


00517 


0941 


omcHi 


0455 


0259 


0(K>49 


0174 


0O882 


00818 


0(1281 


1979 


2455 


1810 


01071 


0769 


00732 


168 


1515 


0865 


1179 


02642 


44(HI 


0273 


0345 


0.2857 


1869 


1176 


0.0545 


0(1755 


02083 


1727 


3190 


03214 


01250 


1220 


170 


2273 


2019 


1014 


3158 


1800 


2455 


01207 


1753 


1589 


0(H)98 


02091 


1774 


1771 


00818 


1724 


2946 


0673 


03537 


172 


0530 


1250 


2069 


0941 


00:00 


3455 


2931 


03312 


012:4 


1824 


0.3909 


02547 


I45S 


0909 


0603 


1339 


0577 


0.2073 


174 


0<J09 


0962 


0517 


0566 


I6II0 


lis: 


0086 


0325 


00935 


(1490 


1656 


2453 


0938 


11 2636 


0690 


0714 


2115 


0732 


mo 


0.0076 


0.0096 


0.0345 


0.0943 


0.0400 


0.0273 


0.0086 


0.0260 


0.0000 


0.0098 


0.0000 


0.0000 


ooono 


0.0000 


0.0000 


0.01 -9 


0.0000 


0.0000 


QaGA2ll9 






































233 


06M2 


0510 


0517 


0064S 


U lOOII 


0424 


0236H 


0395 


0490 


0259 


010-1 


1951 


I.IS3 


0.1071 


0741 


02411 


00119 


0109" 


23S 


0.0IS2 


0.1633 


O.I»97 


0.28'0 


0.2000 


01271 


0.2193 


0.2171 


01961 


0.1293 


0.0446 


0.0976 


0.0745 


0.1071 


0.0556 


0.0625 


0.1667 


0.0488 


241 


4545 


1161 


1276 


47:: 


4SIMI 


2288 


1421 


29<>l 


5784 


5517 


6786 


1291 


4362 


3482 


3889 


2411 


6190 


4390 


243 


143V 


0204 


0690 


0463 


// 0400 


1949 


00439 


144- 


0490 


00S6 


0625 


0012: 


(tOOO 


00000 


0000 


035- 


0000 


0244 


249 


0.0152 


0.1429 


0.0345 


0.0000 


0.0000 


0.1441 


0.0-02 


0.0855 


0.0539 


0.1293 


0.01-9 


0.2683 


0.1064 


0.0804 


O.IIII 


0.2589 


0.166- 


0.1098 


2SI 


0.0303 


O.OOOO 


0.0000 


0.0000 


0.0200 


0.00S5 


0.0000 


0.0132 


0.0392 


0.0086 


0.0000 


0.0000 


0.1277 


0.160- 


0.1296 


0.035- 


0.0000 


0.0244 


255 


00303 


0204 


osf,: 


tmio 


I) 040f> 


O.i.^9 


onoif 


o>y 


0IH4~ 


(torn 


03.^ ~ 


n 0:44 


OOOft 


001-9 


0093 


0000 


00000 


0000 



Differentiation statistics computed over all populations are shown in Table 3. All single-locus estimates 
of among population differentiation (Ost) were found to be significantly different from random 
expectations. Based on stepwise regression analysis, at least one allele at all six of the microsatellite loci 
were found to be significantly (p<0.05) associated with latitude or longitude (see markers in ilalic and 
hold italic, respectively in Table 2). A visual inspection of allele frequencies across the sample sites 
shows a northeast-southwest trend. Allele frequencies tend to be either low in the northeast and high in 
the southwest, or vice versa. Due to this apparent trend, we again performed regression analysis using the 
composite dependent variable (CDV) that combined both latitude and longitude. Results of these 
regression analyses are displayed in Table 4. Frequencies of alleles at all six loci were found to 
significantly vary with the value for CDV. An example of these changes are illustrated in Figure 2 for 
two alleles, one with allele IVequenc\ increasing with CDV distance and the other with allele frequency 
decreasing. Ihe number o I rare alleles per locus was also found to be significantly associated with the 
CDV for three of the six microsatellite loci (Table 4). A visual inspection of the number of rare alleles 
across sample sites again shows a northeast-southwest trend, with higher numbers of rare alleles being 
harbored in southwestern populations. Several loci were also found to be significantly associated w ith 
CDV based on the number of unique alleles. efTecti\c number of alleles, and gene diversity (Table 4). As 
a general trend, there appears to be slighth more alleles and hence more effective numbers of alleles and 



114 



slightly higher levels of gene diversity in southwest populations than in those located in the northeast 
(Figure 2). - _ 



Table 3. Summary of genetic diversity descriptive statistics for six microsatellite 
loci segregating in 18 populations oi Cast an e a dent at a Borkh. located throughout 
the species natural range in eastern North America. 





Sample 














Locus 


Size 


n; 


n. 


h 


ho 


OST 


Nm 


CiCATOl 


1974 


31 


9.222 


0892 


844 


0097 


4655 


C5CAT14 


1974 


15 


3.779 


0.735 


0710 


0,029 


16741 


C5CAT15 


1336 


15 


8.519 


0883 


1 000' 


0.032 


15 125 


OaCA022 


1998 


13 


4198 


0.762 


0.730 


0046 


10370 


OaGA068 


1982 


17 


7.144 


0860 


0786 


0.030 


16 167 


^aGA209 


1936 


15 


4.456 


0.776 


0,705 


0.034 


14206 


Mean 


1870 


17667 


6.220 


0818 


0.755'' 


0.048" 


12.877 


St. Dev 




6653 


2.379 


0.068 


0.059 







^Pa = observed number of alleles, n^ = effective number of alleles, and h = Net's ( 1978) gene diversity, ho = 

observed heterozygosity. <I)st = Michalakis and [ xcoffier's (1996) measure of among population differentiation. 

and Nm = number of migrants exchanged between populations per generation 

''Mean Ost was estimated by summing variance components across loci 

'^observed heterozygosity for this locus was equal to one as the second allele for all trees amplifying only one 

apparent microsatellite allele was scored as unknown or missing data 

''Mean and St Dev do not include ho for locus C5CATI5 



0.25 




• CsCATOl-187 
CsCATOl-192 



SW 



> NE 



250 500 750 1000 1250 
Km alous horizontal 



Figure 2. Plot of allele frequency by composite dependent variable (CDV) expressed in units of 
kilometers along horizontal. 



15 



Table 4. Summar\ of regression analyses for significant associations (Pr>F<0.05) 
between allele frequency, number of rare alleles, observed number of alleles, effective 
number of alleles, and gene diversity and a composite dependent variable (CDV) 
expressed in units of kilometers. 



Frequency 

Locus Allele (bp) 



Regression 



equation 



r.vCATOl 


186 


CvCATOI 


187 


C.sCATOl 


191 


CvCATOI 


192 


r.vCATOl 


200 


r,vCAT14 


145 


CvCATlS 


117 


C.vCATlS 


127 


CvCATlS 


137 


CsCATIS 


139 


(;</CA022 


172 


QciCMll 


174 


^;c/GA068 


164 


faGA068 


180 


QaGMm 


235 


QaGh209 


243 


QaGMW 


249 


QuG/k2Q9 


251 


QaGPaW 


255 


106 


525 


225 


800 


237 


1000 


237 


1250 



Number of Rare Alleles" 



Locus 



Y=0.07264-0 00005024*CDV 

Y=-0.03971+0 00011178*CDV 

Y=-0 01470^0 00006 155*CDV 

Y=0 13460-0 00008108'CDV 

Y=0 08059-0 00004559*CDV 

Y=0 00572-0 00000444*CDV 

Y=0 10464-0 00006849*CDV 

Y=0.05820-0 00003688*CDV 

Y=-0015734 00OO17672*CDV 

Y=0.08369-0 00000003»CDV- 

Y=0 1 9220-0 00000007*CDV- 

Y=0 06282-0 0(X)09 1 97»CDV+O,00000003*CDV- 

Y=0 14814-0 00025546*CDV+0 0000001 1 'CDV- 

Y=0 04318-0 00002958*CDV 

Y=0 1 82 1 3-0 00000005*CDV- 

Y=0. 1 2080-0 000O7540*CDV 

Y=0.0 1 543-0 00008946»CDV 

Y=0 00268-0 00000003*CDV- 

Y=0 04579-0 0000282 1 *CDV 

Y=0 77195-0 0001 1743*CDV 

Y=-0 077 14-0 00033 161 *CDV 

Y=0.8553 1 -0 00008368*CDV 

Y=0 40626-0 00086307*CDV+0 00000037*CDV- 



Regression 
equation 



CvCATOI 
CvCAT15 
QuGMb^ 
All loci 

Obser>ed Number of Alleles'' 

Locus 



Y= 14 18493-0 00490*CDV 
Y=5 13625-0 00331 'CDV 
Y=8.91562-0 01 293*CDV+0 000005 13*CDV- 
Y=58.4O985-007272*CDV+OO0O03079»CDV- 



Regression 
equation 



CvCATOI 
C.VCAT15 
t>aCA068 

Effective Number of Alleles 

Locus 



CsCAT15 
Gene Diversity 

Locus 



Y=0.65276-0 00079907*CDV+0 0O0OO036*CDV- 
Y=0 44260-0 0OO50452*CDV+O0O0O0O22'CDV- 
Y=0.47837-0 00060844*CDV+0 0O0O0O27*CDV- 



Regression 
equation 



Y=8.79012-0 00212*CDV 



Regression 
equation 



CsCAT15 



Y=0 87519-0 00000003 'CDV- 



K' 



0.525 



R^ 



0463 



Pr>F 



0.391 


00055 


0.659 


<0 0001 


0.397 


0051 


0.490 


0012 


0.330 


00127 


0.304 


00177 


0.335 


00118 


0.337 


0115 


0.559 


00004 


0.248 


00353 


0.371 


00073 


0.744 


<0 0001 


0.523 


0.0039 


0.309 


00166 


0.267 


00282 


0.352 


00095 


0.257 


0318 


0.231 


00436 


0.317 


00150 


0.426 


00045 


0.494 


0017 


0.251 


0.0405 


0.624 


0.001 1 



Pr>F 



0.289 


00213 


0.515 


00008 


0.615 


00008 


0.593 


00012 



Pr>F 



0.548 


00026 


0.512 


00046 


0.517 


00043 



Pr>F 



00007 



Pr>F 



0019 



■'number of rare alleles = number of rare alleles in population/number of individuals in population 
""observed number of alleles = number of observed alleles in population/number of individuals in population 



Estimates of genetic distance (D) between pairwise comparisons of populations based on all six loci 
varied from a low of 0.062 to a high of 0.372. averaging 0.206. Similarl\ computed painvisc identity 
estimates ranged from 0.689 to 0.940. \ iclding a mean of 0.8 14. Pairw ise estimates of genetic distance 
were significantly (p=0.001 1 ) associated w ith the geographic distance between paired populations. 
Hov\ever, only a small proportion of the variation found among populations was explained by this 
dependent variable (R'=0.069). Estimates of genetic differentiation (Ost) between pairwise comparisons 
of populations varied from a low of -0.003. to a high of 0.156. and averaged 0.048 across loci. These 
estimates were not significantly associated with geographic distance between the paired populations. 
Thus, populalitms in close geographic pi"o\imit\ tend to have slightl\ higher genetic identities than those 



116 



more geographically distant. Single-locus, as well as multi-locus, UPGMA based on genetic distance 
and PCA based on allele frequencies computed over all sample sites did not reveal patterns of 
differentiation consistent with regional structure. Geographically proximate sample sites did not group 
together, and group membership varied from locus to locus. 



RAPD-based genetic differentiation 

Data describing the RAPD loci used in our analyses are presented in Table 1 . In all populations studied, 
genotypic frequencies observed for microsatellite loci did not significantly deviate from Hardy- Weinberg 
expectations. Assuming then that the RAPD loci we investigated also have genotypes distributed in 
Hardy-Weinberg proportions, we can estimate their allele frequencies from observed frequencies for the 
homozygous null genotypes. Allele frequencies estimated using this approach are displayed by 
population in Table 5. Sixteen of 19 single-locus contingency yj and G" tests for heterogeneity of allele 
frequencies across populations were found to be significant (p<0.05). 



Table 5. Band-present RAPD allele frequencies for 19 loci assayed from samples collected in 17 
populations of Castanea dentata Borkh. located throughout the species natural range in eastern North 
America. Alleles significantly associated with latitude or longitude are identified in italic and bold italic, 
respectively. 



Locus 


CCNC 


BCNC 


GCSC 


PCKV 


RCK1 


SGCVA 


ONTCA 


PCWV 


GCMD 


WCPA 


YCPA 


MPCPA 


LICNY 


RCNY 


HCMA 


MCCT 


ME 


106os()o 


1308 


1722 


0211 


0658 


0426 


1056 


1762 


2929 


1728 


0839 


0887 


0780 


2421 


1220 


1982 


0364 


0917 


196os2s 


0.31)114 


0.I6II 


01340 


02138 


01584 


01244 


0.1136 


0.1921 


0.0547 


01036 


O0887 


00780 


00324 


01835 


0.0000 


00000 


01056 


106ot,5o 


7446 


6400 


8000 


8652 


5918 


1 0000 


8093 


8830 


06181 


6220 


8159 


8419 


7083 


7980 


08110 


7327 


7261 


106o7(jo 


0000 


0474 


1835 


00187 


0646 


00513 


0090 


00138 


0000 


0009 


0087 


0382 


0435 


0000 


0272 


0364 


0126 


1 0^0800 


7051 


1 0000 


1 0000 


1 0000 


1 0000 


1 0000 


1 0000 


8821 


1 0000 


6727 


08143 


8419 


1 0000 


7474 


1 0000 


1 0000 


1 0000 


1^-^0450 


0.022 J 


O0973 


0000 


o.onvo 


0000 


0433 


0001) 


0344 


o.oono 


0000 


0000 


o.ooon 


0000 


0000 


00S7 


0.0000 


0.0126 


184,150 


0435 


0973 


1340 


0646 


1056 


0980 


0770 


0859 


0786 


0299 


0691 


0000 


0000 


02175 


0087 


0000 


00126 


184|((,„i 


1495 


3061 


1340 


1443 


1835 


3622 


2230 


2289 


3064 


1034 


2042 


2929 


3386 


0632 


1982 


3453 


3481 


2I3omi 


0.2421 


01368 


0.2929 


01798 


0911 


02254 


01972 


01S35 


01815 


01679 


O0728 


0.0968 


02421 


03353 


06526 


1. 0000 


02745 


2\hm> 


3477 


4059 


4084 


4606 


4599 


2362 


02112 


3280 


3660 


2279 


2867 


05412 


4947 


2462 


2689 


1 0000 


3206 


225„„ 


0.7446 


1.0000 


1.0000 


1.0000 


1.0000 


1 0000 


1.0000 


08821 


ft 7621 


0.6331 


0.4702 


0.5918 


O6703 


04161 


0.5044 


08333 


0.6508 


225i45o 


0022 


0000 


0000 


0000 


0000 


0000 


1 0000 


0000 


0000 


0000 


0000 


0000 


0000 


0000 


0000 


0897 


0000 


237„,,5 


3386 


2672 


2546 


04117 


02517 


2362 


1310 


1780 


2327 


2279 


1943 


4689 


1340 


0426 


3841 


1590 


2510 


237,^ 


0.0871 


00093 


0.1416 


01611 


02789 


0.0084 


00417 


0.0921 


0.0483 


01121) 


0.0177 


0.0129 


O087I 


0.0105 


00262 


0.0000 


0.0247 


237,2,0 


30H4 


1)3914 


0S51I 


2302 


25jy 


1633 


0632 


3175 


1035 


01914 


0S23 


owl 


1)0211 


1:4s 


1906 


o3h: 


2510 


423„»„ 


1403 


1835 


0426 


1158 


0835 


1633 


1313 


2536 


02341 


2352 


2277 


0267 


1882 


3773 


0903 


2632 


1377 


423„„5 


00109 


0392 


0675 


0180 


0408 


0426 


0190 


0598 


0652 


0801 


0267 


0823 


00114 


0742 


0267 


0000 


0392 


500„T75 


0000 


0000 


0000 


0000 


0000 


0084 


0000 


0066 


0246 


0000 


0000 


0000 


0572 


0000 


0457 


0000 


0123 


5140575 


1972 


3412 


3406 


2799 


1282 


1734 


2459 


2494 


03614 


03140 


Gim 


1029 


0585 


5337 


6026 


2735 


3753 



Differentiation statistics computed over all populations are presented in Table 6. Estimates of among 
population differentiation (Gst) were found to be significantly greater than zero at 14 of the 19 loci. 
Based on our stepwise regression analysis, allele frequencies at six of the 19 RAPD loci were 
significantly (p<0.05) associated with latitude or longitude (markers in italic and boid-ita/ic, respectively 
in Table 5). As was observed for the microsatellite loci, a visual inspection of allele frequencies across 
the sample sites showed a northeast-southwest trend. Again, we performed regression analysis using the 
CDV. Four loci were found to be significantly associated with the CDV (Table 4). At all four loci, band- 
present allele frequencies were higher in southwest populations than in those from the northeast. 

Estimates of genetic distance (D) between pairwise combinations of populations computed across loci 
varied from a low of 0.003, to a high of 0. 144, with a mean value of 0.037. Similarly computed pairwise 
identity estimates ranged from 0.866 to 0.997, with a mean of 0.964. Unlike the microsatellite data. 



117 



pairwise estimates of genetic distance were not significantly (p=0.0571) associated with the geographic 
distance between paired populations and neither were pairwise estimates of genetic differentiation. As 
was the case for the microsatellite loci, single-locus or multi-locus UPGMA computed from RAPD 
genetic distances, or PCA based on RAPD allele frequencies, did not reveal differentiation patterns 
suggestive of regional structure. 



Table 6. Summary of genetic diversit>' descriptive statistics for 19 RAPD loci 
assayed from samples collected in 1 7 populations of Caslanea dentala Borkh. 
located throughout the species natural range in eastern North America. 



Sample 
Locus Size a. h Gst Nm 



1 06o500 
106o525 
1 06o65(| 
1 06(|7(K) 

1 06(181111 

184„450 

184„5„ 
184 18011 

^ I 3i)»j(m 

2 1 3 1000 
225o8oo 
225 1450 
237(1825 

^J /lodo 

237i25(i 
423(Kioo 

423(1875 
500„775 
514,1575 

Mean 
St. Dev 



845 


1.321 


0.243 


0.049 


9.802 


845 


1.273 


0.214 


0.056 


8443 


849 


1 552 


0.356 


0.062 


7.628 


844 


1 051 


0.049 


0056 


8.471 


843 


1.178 


0.I5I 


184 


2.222 


883 


1 030 


0.029 


0.046 


10.364 


881 


1 142 


0.124 


0.050 


9413 


878 


1 567 


0.362 


0.047 


10.080 


794 


1 535 


0.348 


0067 


6914 


801 


1.818 


0.4.50 


-0.006 


2000.0 


808 


1 598 


0.374 


-0.008 


2000.0 


810 


1.010 


0010 


-0.336 


2000.0 


871 


1.578 


0366 


0060 


7.863 


873 


1 126 


112 


081 


5.709 


869 


1 416 


0294 


0081 


5.642 


861 


1 426 


0299 


0,053 


8889 


858 


1 090 


0,082 


0.016 


30.536 


870 


1 021 


0.021 


031 


1 5.406 


858 


1.700 


414 


093 


4 899 


850 


1 339 
0.258 


0226 
0.148 


0.036 


9.517" 



'^ric = effective number of alleles, and h = Nei's (1978) gene diversity. Gst = Nei's (1987) measure of among 
population differentiation, and Nm = number of migrants exchanged between populations per generation 
"Mean excludes estmiates for loci 213|(»h,, 225(is(hi- and 225 145 



DISCUSSION 

One of our main concerns in this investigation was inclusion of trees that are not pure American chestnut. 
Inappropriate trees include interspecific hybrids or pure species other than American chestnut, especiallv 
the native congener species chinkapin {Caslanea piiwila). Inclusion of such contaminants could have 
inflated our estimates of genetic diversity, especially in populations containing the non-American 
chestnut samples, as well as clouded true patterns of genetic variability. Chloroplast DNA sequence 
variations have been widely used to investigate interspecific relationships among plant species (Palmer et 
al. 1988, Clegg et al. 1991) because they evolve slowly. We identified a chloroplast-speciflc marker 
(primers a and b: Taberlet et al. 1991 ) that quicklv differentiates American chestnut chloroplast DNA 
from all other Caslanea species, including the native ('. piiniila. Unfortunatelv. maternal inheritance of 
chloroplasts precludes our abilit> to distinguish interspecific hybrids of maternal American chestnut 
origin. As a result, our sample set might still contain some interspecific hybrids, however, the number 
should be small as most collections were made in either State Forests or National Forests where non- 
native Caslaiwa species do not extensively occur. 



18 



Our results demonstrate that high levels of microsatellite and RAPD variability exist in American 
chestnut, and that most of this variation occurs within local populations (95.2% and 96.4%, respectively). 
These results are comparable to observations made in other long-lived, outcrossing, woody plant species 
(Hamrick and Godt 1990; Hamrick et al. 1992), where as a rule, greater than 90% of the variation occurs 
within populations. Our results are also consistent with previous observations of allozyme variability in 
C. sativa and American chestnut, where 90% of the diversity was reported to exist within populations 
(Pigliucci et al. 1990; Huang et al. 1998). Whereas only scant evidence for a cline in allele frequency 
variation (alleles at 1 of 14 polymorphic allozyme loci) was previously reported for American chestnut 
(Huang et al. 1998), our results clearly demonstrate that a cline in allele frequencies and number of rare 
alleles exists along the Appalachian axis. Clinal variation of allele frequencies along latitudinal and 
longitudinal gradients has been reported for a number of tree species (Lagercrantz and Ryman 1990; 
Zanetto and Kremer; Leonardi and Menozzi 1995, Tomaru et al. 1997), including C. sativa (Pigliucci et 
al. 1990; Villanietal. 1991; Villani et al. 1992; Villani et al. 1994). The main proposition set forth to 
explain this phenomenon is that geographical variation in allele frequencies resulted from post-glacial 
migration and founding events. Such processes are consistent with the patterns of variability we observed 
for American chestnut. The highest levels of gene diversity and the greatest numbers of rare alleles are 
found in the southwestern portion of its range. This suggests that its glacial refugium existed in the 
southeastern U.S.. perhaps extending southward into the Gulf Coastal plain of present day Mississippi and 
Alabama. As a general finding, American chestnut still exists as a highly variable species, even at the 
margins of its natural range, with a large proportion of its genetic variability occurring within populations. 
Furthermore, existence of the clinal pattern of variation implies that extensive gene flow took place 
among populations before the spread of chestnut blight. 

Although most of the genetic variation found in American chestnut occurs within local populations, a 
statistically significant proportion exists among populations. Magnitudes of the <i>sj and Gsi estimates 
obtained in our investigation are slightly lower than those reported for American chestnut by Huang et al. 
(1998). In this research we used a chloroplast-specific marker to identify trees that were not pure 
American chestnut and excluded these individuals. However, Huang et al. (1998) did not take 
precautionary measures to identify aberrant specimens. Inclusion of such individuals in some samples 
will tend to inflate levels of among population differentiation. Although our estimates of among 
population differentiation might be considered low, Osr values obtained for all six microsatellite loci and 
GsT values obtained for 14 of the 19 RAPD loci studied indicate that populations significantly differ in 
allele frequency. Moreover, population pairwise estimates of genetic distance, based on microsatellite 
haplotype frequencies, were shown to be significantly associated with the geographic distance between 
populations. Thus we conclude that geographically proximate populations are slightly more genetically 
similar than geographically distant populations. These findings lead us to conclude that although long 
distance gene flow was possible in the past, it was infrequent enough to allow genetic differentiation to 
take place. 

From UPGMA and PC A analyses, it is evident that regional differentiation did not occur in American 
chestnut. Geographically proximal populations did not group together, and group make-up differed 
across loci. In contrast, Huang et al. (1998) concluded that a somewhat weak and incomplete pattern of 
regional differentiation exists, based largely on latitudinal differences. Although the results obtained by 
UPGMA and PCA of the allozyme data were interpreted as being suggestive of regional structure, 
hypothetical regional effects were not quantified by means of a hierarchical AMOVA, nor were statistical 
tests employed to detect differences. Because of our more comprehensive sampling of the natural range 
(18 populations versus 12), larger sample sizes collected (average 55 trees per population versus 22 trees), 
and elimination of suspect samples (i.e. trees that did not have the characteristic American chestnut 
chloroplast haplotype), we believe the results obtained in this investigation represent a more accurate 
picture of population structure in American chestnut. 



19 



Our findings clearly demonstrate that American chestnut still exists as a highly variable species 
throughout its entire native range. In spite of this high variabilit>. we must point out that the results of 
this study represent variability existing in the pre-blighted forest, and caution that unless measures are 
taken to restore American chestnut and enhance opportunities for it to sexually reproduce, this species 
will likely face serious erosion of its gene pool as root systems fail to produce sprouts and die. Along 
these lines, results of this study can be used as a baseline in the future for assessing the degree and 
rapidity of such a decline. 

Taking into account the differentiation observed at these loci, no disjunct regional pattern of variation 
exists. Prior to introduction of the blight, genetic variability in American chestnut followed a pattern 
consistent with the hypothesis of a single metapopulation where genetic drift played a major evolutionary 
role. Currently, approximately 95% of the neutral genetic variation of the species can be captured by 
sampling within any one population. However, the results of this study are based on neutral genetic loci 
and do not necessarily reflect genetic differentiation for adaptive genes or gene complexes. Therefore, in 
order to assure that most of the variation produced by these genes is also captured in conservation and 
breeding endeavors, sampling should focus on collecting a fairly large number of individuals (50 to 100 
or more) from each of several geographic areas. 



LITERATURE CITED 

Barr, M.E. 1979. The diaporthales in North America with emphasis on gnomonia and its segregants. 
Mycologia Memoir 7, J. Cramer, Lehre, Germany. 

Bumham, C.R. 1981. Blight-resistant American chestnut: There's hope. Plant Dis. 65:459-460. 

Clapper, R.B. 1954. Chestnut breeding techniques and results. II. Inheritance of characters, breeding for 
vigor, and mutations. J. Hered. 45:201-208. 

Clegg, M.T., G.H. Learn, and E.M. Goldberg. 1991. Molecular evolution of chloroplast DNA. P. 135-149. in 
Evolution at the molecular level, Selander. R.K., A.G. Clark, and T.S. Whittam (eds.) Sinauer Associates Inc., 
Sunderland, MA. 

Excoffier, L., and M. Slatkin. 1995. Maximum-likelihood estimation of molecular haplot>pe frequencies 
in a diploid population. Mol. Biol. Evol. 12:921-927. 

Hamrick, J.L., and M.J. Godt. 1990. Allozyme diversit> in plant species. P. 43-63. in Plant population 
genetics, breeding, and genetic resources. Brown, H.D., M.T. Clegg, A.L. Kahler, and B. Weir (eds.). 
Sinauer Associates Inc., Sunderland. MA. 

Hamrick, J.L.. M.J. Godt, and S.L. Shemian-Broyles. 1992. Factors influencing levels of genetic diversity in 
woody plant species. P. 95-124. in Population genetics of forest trees. Adams. W.T.. S.H. Strauss. D.L. Copes, 
and A.R. Griffin (eds.). Kluwer Academic Publishers. London. 

Hebard, F.V. 1994. The American Chestnut Foundation breeding plan: beginning and intemiediate steps. J. Am. 
Chestnut Found. 8(1 ):2 1 -28. 

Huang. H.. F. Dane, and T.L. Kubisiak. 1998. Allozyme and RAPD anal\sis of the genetic diversity and 
geographic variation in wild populations of the American chestnut (Fagaceae). Am. J. Hot. 85:1013-1021. 



120 



Kubisiak, T.L., F.V. Hebard, CD. Nelson. J. Zhang, R. Bematzky, H. Huang, S.L. Anagnostakis, and 
R.L. Doudrick. 1997. Molecular mapping of resistance to blight in an interspecific cross in the genus 
Castanea. Phytopathology 87:751-759. 

Lagercrantz, U., and N. Ryman. 1990. Genetic structure of Norway spruce {Picea abies): concordance of 
morphological and allozyme variation. Evolution 44:38-53. 

Leonardi, S., and P. Menozzi. 1995. Genetic variability of Fagus syhatica L. in Italy: the role of 
postglacial recolonization. Heredity 75:35-44. 

Marinoni, D., A. Akkak, G. Bounous, K. Edwards, and R. Botta. 2003. Development and characterization 
of microsatellite markers in Castanea sativa {M'\\\.). Mol. Breed. 1 1(2):127-136. 

Michalakis, Y., and L. Excoffler. 1996. A generic estimation of population subdivision using distances 
between alleles with special reference for microsatellite loci. Genetics 142:1061-1064. 

Nei, M. 1972. Genetic distance between populations. Am. Natur. 106:283-292. 

Nei, M. 1978. Estimation of average heterozygosity and genetic distance from a small number of 
individuals. Genetics 89:583-590. 

Nei, M. 1987. Analysis of gene diversity in subdivided populations. P. 187-192. in Molecular 
Evolutionary Genetics. Columbia University Press, New York. 

Palmer, J.D.. R.K. Jansen, H.J. Michaels. M.W. Chase, and J.R. Manbart. 1988. Chloroplast DNA 
variation and plant phylogeny. Ann. Miss. Bot. Gar. 75:1 180-1206. 

Pigliucci, M., S. Benedettelli, and F. Villani. 1990. Spatial patterns of genetic variability in Italian 
chestnut {Castanea sativa). Can. J. Bot. 68:1962-1967. 

Sargent, C.S. 1905. Manual of trees of North America. Houghton Mifflin Co., Boston and New York. p. 

220-222. 

SAS Institute Inc. 1999. SAS/STAT Users Guide, Version 8. SAS Institute, Inc., Cary, NC. 3884 p. 

Schneider. S., D. Roessli, and L. Excoffier. 2000. A software for population genetics data analysis. 
Genetics and Biometry Laboratory. University of Geneva. Switzerland. 

Stephenson, S.L., H.S. Adams, and M.L. Lipford. 1991. The present distribution of chestnut in the upland 
forest communities of Virginia. Bull. Torrey Bot. Club 1 18:24-32. 

Taberlet, P., L. Gielly. G. Patou. and J. Bouvet. 1991. Universal primers for amplification of three non- 
coding regions of chloroplast DNA. Plant Mol. Biol. 17:1 105-1 109. 

Tomaru. N., T. Mitsutsuji, M. Takahashi. Y. Tsumura, K. Uchida, and K. Ohba. 1997. Genetic diversity 
in Fagus crenata (Japanese beech): influence of the distributional shift during the late-Quaternary. 
Heredity 78:241-251. 

U.S. Census Bureau. 1908. The lumber cut of the United States, 1907. For. Products 2:1-53. 



121 



Villani, F., M. Pigliucci, S. Bcnedettelli, and M. Cherubini. 1991 Genetic differentiation among Turkish 
chestnut {Castanea sativa Mill.) populations. Heredity 66: 131-136. 

Villani, F., M. Pigliucci, and M. Cherubini. 1994. Evolution of Castanea sativa Mill, in Turkey and 
Europe. Genetic Research, Cambridge. 63: 109-166. 

Villani, F., M. Pigliucci, M. Lauteri, and M. Cherubini. 1992. Congruence between genetic, " 

morphometric, and physiological data on differentiation of Turkish chestnut {Castanea sativa). Genome 

35:251-256. 

Weir, B.S. 1990. Methods for discrete population genetic data. P. 110-113 in Genetic data analysis. ^ 
Sinauer Associates Inc., Sunderland, Mass. 

Yeh. F.C., R-C. Yang, T.J. Boyle. Z-H. Ye, and J.X. Mao. 1997. POPGENE, the user-friendly shareware 
for population genetic analysis. Molecular Biology and Biotechnology Centre, University of Alberta, 
Canada. 

Zanetto, A., and A. Kremer. 1995. Geographical structure of gene diversity in Quercus petraea (Matt.) 
Liebl. \. Monolocus patterns of variation. Heredity 75:506-517. 



ACKNOWLEDGEMENTS 

The authors thank Glen Johnson and Charles Burdine for their skilled technical assistance with DNA 
extraction. PCR amplification, microsatellite marker detection, and data acquisition: Andv Dav id and 
Dave Wagner for the white oak microsatellite primer sequences; Gabriele Baccaro and Roberto Botta for 
releasing the European chestnut microsatellite primer sequences prior to publication of their manuscript: 
and Sandra Anagnostakis; Dave Annstrong, Glen Beaver. Robert Bernatzky; Mary Bunch. Peter Carson, 
Hill Craddock, Mark Double, Fred Hebard, Craig Hibben, E. Kenneth James, Michael Kluempke, Jeff 
Lewis, Paul Sisco. Bob Summersgill, Wayne Swank. Melissa Thomas-VanGundv. Wells Thurber. Cathy 
Townsend, Stan Webb and Eric Weisse for their assistance in collecting chestnut samples for this study. 



1 Am^m 



Steiner, K. C. and Carlson. J. E, eds. 2006. Restoration of American Chestnut To Forest Lands - Proceedings of a 
Conference and Workshop. May 4-6, 2004, The North Carohna Arboretum. Natural Resources Report 
NPS/NCR/CUE/NRR - 2006/001, National Park Service. Washington, DC. 

REGIONAL ADAPTATION IN AMERICAN CHESTNUT 

;T_-j, Kim C. Steiner 

N^ ' School of Forest Resources, The Pennsylvania State University, 

' ^ University Park, PA 16802 (kcs@psu.edu) 



Abstract: Conservation of forest genetic resources, such as restoration of American chestnut, requires 
knowledge of genetic variation patterns in adaptive, non-neutral alleles. Almost no such infonnation is 
available for American chestnut, but there is information available from other forest tree species including 
species that are sympatric with chestnut. Some of that information is summarized in this paper, and the 
adaptive significance of several growth and physiological characteristics is discussed. Most tree species 
exhibit "racial" patterns of genetic variation that parallel geographic gradients in climate. Wild 
populations that have survived in a locality for many generations have a genetic identity of place that 
reflects a history of natural selection and adaptation. Judging from genetic variation patterns in sympatric 
species, American chestnut populations are probably genetically distinct in important and somewhat 
predictable ways. American chestnut breeding and restoration projects should be guided by this 
knowledge. 

Keywords: genetic / geographic / racial / variation / selection / adaptation / growth rate / cold tolerance / 
phenology 



INTRODUCTION 

With new technologies for controlling chestnut blight on the horizon, we are beginning to contemplate the 
restoration of American chestnut to something like its former importance in our forests. The creation of 
genetically resistant trees through breeding or genetic engineering is especially promising as a foundation 
for restoration efforts. However, blight-resistant alleles can be introduced to (or found in) only a tiny 
fraction of the chestnut trees that still survive, so any restoration program that employs blight-resistant 
trees will inevitably force the species through a genetic '"bottleneck" with the danger that important alleles 
may inadvertently be lost. Like other plant species, especially those with large natural distributions, 
American chestnut undoubtedly contains a great deal of genetx variation. This diversity should be 
protected (Irwin 2003), and indeed it should be exploited if possible in the restoration process itself. But 
to do so will require an understanding of how genetic variation is structured within the species. 

Infonnation about range-wide genetic variation in American chestnut is limited to studies of allozymes 
and DNA markers of unknown and probably neutral adaptive significance (Huang et al. 1998. Kubisiak 
and Roberds 2003). These studies revealed that differences among populations account for only 5 to 10 
percent of the total genetic variation measured, results that closely resemble the findings of other studies 
of neutral alleles in species that are similar to American chestnut in mating system, longevity, population 
size and density, and other characteristics that affect gene flow (Hamrick and Godt 1990). This is useful 
knowledge - for example, it tells us that gene flow among populations has been relatively strong within 
this species - but it tells us virtually nothing about the structure of genetic variation in alleles subjected to 
the pressures of natural selection. Patterns of variation in adaptively neutral genetic markers may bear 
little relationship to patterns of variation in adaptively relevant alleles, whose variation patterns may also 
differ from one another according to what kind of characteristics they control (Morgenstern 1996). 
Genetic variation in "fitness" characteristics (in the terminology of Darwinian theory), especially those 
responding to regional selection gradients such as climate, is highly relevant to conservation or restoration 
efforts undertaken at a regional or range-wide scale. 

123 



Unfoilunately. we know almost nothing about genetic variation in fitness characteristics within American 
chcslnul. There is, however, a fairly substantial body of such infonnation from studies of other tree 
species, and this knowledge can be used to inform future decisions about American chestnut, fhis ^" 
literature comes from replicated, "provenance" tests of progeny from natural populations grown in a 
common environment. Such experiments, if properly designed, permit the researcher to apply the 
methods of quantitative genetics to measure and test contributions of genetic variation to phenot> pic 
variation, even without knowing the underlying DNA structure or mode of gene action. It is even 
possible to partition the relative contributions of among- versus within-population genetic variance just as 
population geneticists do when working directly with DNA markers. In this paper I pro\ ide some 
examples of such research on eastern forest tree species. Although we cannot know for certain, American 
chestnut would probably exhibit similar genetic variation over similar environmental gradients if it were 
studied in the same way. 



EXAMPLES OF REGIONAL VARIATION IN ADAPTIVE CHARACTERISTICS 

Bud-Burst Timing in Eastern American Species 

Deciduous trees can take advantage of abundant moisture and shade-free conditions (in the case of plants 
growing below the forest canopy) by initiating growth early in the spring, but early growth initiation 
increases the risk of frost injur>' to young leaves. Thus, genetic control of bud-burst timing is probably 
under strong selective pressure for optimality in any given environment. It is topically the case in 
common garden tests that populations from colder (more northern or higher ele\ alion) en\ ironments burst 
bud earlier in the spring. Bud burst in trees is usually cued by rising temperature, and populations from 
colder climates are adapted to grow (and begin growing) under cooler temperatures. When the 
geographic pattern of bud-burst timing is the opposite {e.g.. southern populations earlier in common 
gardens), as has been recorded in a few species, it is likely attributable to a different environmental cue 
for growth (photoperiod) rather than a fundamental difference in the way the plant has adapted to 
environmental gradients (Steiner 1979a). Of course, in nature, plants in warmer climates alwa\s begin 
growing before plants in colder climates. But if the onset of spring is heralded b\ the appearance of 
leaves on trees, spring would be even more delayed in the north if all populations of a species required 
equally warm temperatures for growth. 

Steiner (1975 and 1979b) described geographic patterns of genetic variation in bud-burst timing in three 
species that are broadly sympatric with American chestnut: yellow birch, eastern white pine, and Virginia 
pine. The first two species are sympatric w ith chestnut throughout most of its Appalachian distribution, 
but they also occur widely in the Lake Stales and southeastern Canada. Virginia pine occurs naturally 
only from central Pennsylvania southward, but the whole of its distribution is very similar to the southern 
half of American chestnut's. All three species showed the typical north-early / south-late pattern o'i 
variation in bud-burst timing in common-garden tests. Also, all three exhibited genetic \ariation within 
their area of sympatr\ with chestnut, with clinal gradients statisticalK detectable o\er minimum distances 
vi'i 100 to 300 km. A partial exception was yellow birch, which showed wo clear latitudinal gradient in 
bud-burst timing in from Pennsylvania southward (roughly the southern half of the American chestnut 
range). Steiner (1975) also found that population variation in time of flowering (pollen release) generally 
corresponded with population variation in bud-burst timing. This may not be the case with the late- 
flowering American chestnut, whose habit of ilowcring in mid-summer ma\ not be so closeh linked in 
both a physiological and genetic sense with its phenolog\ of \egetati\c growth. 

Broad climatic gradients in genetic variation are almost always somewhat muddled by populations that do 
not fit the trend, and this was true in Steiner's studies. Elevations of oriain differed ureatlv for the 



124 



Appalachian populations of all three species, and it is reasonable to suppose that adaptation to elevation 
might have explained locally "anomalous" populations. However, there was no detectable relationship 
between elevation and bud-burst timing, at least after accounting for latitude (the more southern 
populations tended to occur at higher elevations). A better test of the effect of elevation of origin on 
genetic variation would be to deliberately sample a number of populations along an elevational transect 
up and down a single mountain. I am not aware of any such study in the Appalachians, but McGee 
(1974) studied four populations of northern red oak collected from different elevations "within 100 km of 
Asheville, North Carolina" and found a possible elevational effect on genetic differentiation in bud-burst 
timing in that species. From these studies we can predict that American chestnut seed that is moved to 
environments that are warmer or colder than the native environment will likely be somewhat "off" in the 
timing of new growth in the spring, growth cessation in late summer, and perhaps flowering. 



Genetic Variation in Cold Tolerance in Two Species 

The process of acclimation to cold in woody plants begins after the cessation of growth and is triggered 
by diminishing day length. Acclimation deepens as plants experience increasingly colder temperatures, 
reaching a maximum when temperatures are coldest, in January or February. This process has a 
metabolic cost, and plants that develop greater levels of cold tolerance presumably pay a price for that 
advantage. The benefit to a plant of adequate cold tolerance, and the cost of its absence, is clear and 
direct. Not surprisingly, geographic patterns of genetic variation in ability to acclimate to low 
temperature tend to look very much like January low-temperature isotherms on a map. 

Williams (1984) described a nicely done study of genetic variation in cold tolerance within green ash. 
Green ash has a native range that extends far beyond the region in which American chestnut grows, but 
Williams" study included seven populations of green ash from the area of overlap with our spec'es of 
interest. These populations differed in the expected fashion: the rapidity of acclimation and depth of 
mid-winter cold tolerance were greatest in New York and Pennsylvania populations, intermediate in 
central Virginia populations, and least in eastern Tennessee populations. These represented nearly half of 
the total range of variation in mid-winter cold tolerance levels for all green ash populations studied 
(which included Manitoba and South Dakota provenances, but none more southern than Tennessee). 
Williams also found significant M77/7/>?-population genetic variation in cold tolerance - except in 
populations near the northern limit of the range, where the species is presumably at its limit of adaptation 
to cold. 

The results described above were obtained under controlled, laboratory conditions using twigs taken from 
trees grown out-of-doors. Williams (1984) also measured actual winter injury over a three-year period in 
nine replicate plantings of a green ash provenance test in the upper Midwest and Northeast. As one 
would expect, trees that had originated from progressively wanner climates had progressively more 
severe winter injury. Put another way, the fraction of trees that escaped winter injury diminished as 
population origin went from north to south. Interestingly, Williams found that some trees escaped injury 
when growing in environments colder than where they originated, sometimes even much colder, but 
moving a population to colder environments was always accompanied by an increase in the risk of winter 
injury as measured by percentage of trees injured. 

The mating systems, life history characteristics, and population structure of American chestnut and many 
other forest tree species tend to promote gene flow, which acts to minimize genetic differentiation 
between nearby populations (Loveless and Hamrick 1984). There are very few known instances of 
clearly adaptive genetic differentiation in temperate forest trees occurring over distances of a few 
kilometers or less. One example is that described by Berrang and Steiner (1986) and Steiner and Berrang 
(1990) for cold tolerance variation in pitch pine. Pitch pine and American chestnut have almost identical 



125 



distributions and often co-occur on the same sites. Near State College, Pennsylvania, pitch pine is 
common within an area called the "Barrens," which often (and during any month of the year) has 
substantially lower nighttime temperatures than the surrounding countryside. Pitch pine also grows in 
areas near the Barrens that have locally normal temperatures. These authors compared the development 
of cold tolerance in dormant seedlings raised under controlled conditions but originating from Barrens 
and non-Barrens trees. Compared to the neighboring population on normal sites. Barrens seedlings 
acclimated more rapidly in the fall, achieved greater levels of cold tolerance in mid-winter, and de- 
acclimated more slowly in the spring. Evidently, differences in selection pressure over the distances 
separating these populations (about 8 km) have been strong enough to create genetically distinct 
populations. 



Genetic Variation in F4eiaht Growlh Rate in Northern Red Oak 

Everyone knows that plant growlh is greatly influenced by environmental conditions, but growlh rate is 
always under genetic control, as well. Growth rate is fundamentally important to plant fitness, though 
rapid growth is not always (or even usually) the best strategy for ecological success. As Grime ( 1 979) 
has pointed out, there is essentially a "zero-sum" relationship between plant investment in growth, 
reproduction, and defense or toleration of stress - the benefits vary according to circumstances, but the 
costs are always there, and a plant cannot afford to excel at ever\ thing. In range-w ide provenance tests of 
species whose distributions span warm/cold or wet/dr> climatic gradients, it is t>picall\ the case that 
populations from wanner or moister environments are genetically capable of faster growth than is found 
in their poor relations living at the limits of hospitable conditions (Wright 1976). This pattern probably 
arises because competition is a stronger selective force in the milder climates (favoring rapid growth), and 
stress is a stronger force in the harsher environments, where investment in cold tolerance or drought 
tolerance or avoidance (always at some metabolic cost) is more advantageous than producing more 
foliage. 

However, this generalization applies most particularly to species that inhabit a truly wide range of 
environments. Through the smaller and more homogenous region occupied b> American chestnut, 
patterns of genetic variation in growth rate have typicalK had a strongly "random" character in 
provenance tests, usually defying simple geographic description. The occurrence of apparently random 
variation does not necessarily mean that the controlling alleles are selectively neutral, but it may mean 
that the forces that favor or disfavor rapid growth are more local than regional in occurrence. 

Northern red oak occupies the same region as American chestnut plus a little more, occurring from the 
Gulf Coastal Plain (but not Florida) to eastern Kansas and southern Ontario. Several genetic tests of this 
species (summarized and reanalyzed by Steiner 1998) were designed to permit the partitioning of genetic 
variation into "local" and "regional" components. In tests that included populations no more than a few 
hundred kilometers apart, virtually all of the genetic variation in growth rate occurred within populations 
(populations differed little or not at all). However, even in a "range-wide" test (with distances between 
populations of up to 2000 km), vvithin-population genetic variance still accounted for 64 percent of total 
genetic variance in growth rate. Furthermore, variation ainon}^ noithern red oak populations in growth 
rate, when it occurs, does not show a clear and straightforward geographic pattern (Steiner 1998). Steiner 
argued that selective forces acting on growth rate in northern red oak may plausibK operate on a very 
local, "microsite" scale in stands of this species. (Forest soils are t\'picalK verv heterogeneous vis-a-vis 
growth potential in northern red oak.) If American chestnut has a similar pattern of genetic variation, 
then a tvpical population mav contain most of the genetic variation that occurs within the species in 
alleles controlling growth rate. However, the possibilit> of important, inter-population variation should 
not be ignored. 



126 



^^ .^ CONCLUSIONS 

We know almost nothing about genetic variation in characteristics of American chestnut that play a role 
in adaptation to the environment. However, most tree species exhibit "racial" patterns of genetic 
variation that parallel geographic gradients in climate. Wild populations that have survived in a locality 
for many generations have a genetic identity of place that reflects a history of natural selection and 
adaptation. When environmental differences are large enough, natural selection may favor genetic 
differentiation even on a rather local scale. Studies of other species have consistently revealed genetic 
variation - within the species' region of sympatry with American chestnut - in adaptively important 
characteristics such as growth rate, phenology, and cold tolerance. Disrupting these variation patterns by 
careless human meddling can result in trees that are unsuited to their environments in subtle but perhaps 
important ways, particularly considering that trees with nonnal lifespans must survive many decades of 
environmental vicissitude. In the absence of evidence to the contrary, we should expect that American 
chestnut populations also differ genetically from one another in similar ways. This knowledge should 
guide breeding and restoration projects in American chestnut. Restoration projects should seek to 
preserve as much natural genetic variation as possible within American chestnut, and blight-resistant trees 
used to restore wild populations should be derived from locally or regionally native American chestnut 
trees. 



LITERATURE CITED 

Berrang, P.C., and K.C. Steiner. 1986. Seasonal changes in the cold tolerance of pitch pine. Can. J. For. 
Res. 16:408-410. 

Grime, J. P. 1979. Plant strategies and vegetation processes. John Wiley and Sons, New York. 222 p. 

Hamrick, J.L., and J.W. Godt. 1990. Allozyme diversity in plant species. P. 43-63 in Plant Population 
Genetics, Breeding, and Genetic Resources, Brown, A.H.D. et al. (eds.). Sinauer Assoc. Inc., Sunderland, 
MA. 

Huang, H., F. Dane, and T.L. Kubisiak. 1998. Allozyme and RAPD analysis of the genetic diversity and 
geographic variation in wild populations of the American chestnut (Fagaceae). Am. J. Bot. 85:1013-1021. 

Irwin, H. 2003. The road to American chestnut restoration. J. Amer. Chestnut Found. 16(2):6-]3. 

Kubisiak, T.L., and J.H. Roberds. 2003. Genetic variation in natural populations of American chestnut. J. 
Am. Chestnut Found. 16(2):42-48. 

Loveless, M.D., and J.L. Hamrick. 1984. Ecological determinants of genetic structure in plant 
populations. Ann. Rev. Ecol. Syst. 15:65-95. 

McGee, C.E. 1974. Elevation of seed sources and planting sites affects phenology and development of red 
oak seedlings. For. Sci. 20:160-164. 

Morgenstem, E.K. 1996. Geographic variation in forest trees. University of British Columbia Press, 
Vancouver. 209 p. 

Steiner, K.C. 1975. Patterns of genetic variation within fifteen trees species in times of bud burst and 
flowering. Ph.D. thesis, Michigan State University, 170 p. 



127 



Steiner. K.C. 1979a. Variation in bud-burst timing among populations of interior Douglas-fir. Silv.^ ^ , 
Genet. 28:76-79. _ -I 

Steiner, K.C. 1979b. Patterns of variation in bud-burst timing among populations in several Pinus species. 
Silv. Genet. 28:173-256. 

Steiner. K.C. 1998. A decline-model interpretation of genetic and habitat structure in oak populations and 
its implications for silviculture. Eur. J. For. Path. 28:1 13-120. 

Steiner, K.C. and P.C. Berrang. 1990. Microgeographic adaptation to temperature in pitch pine progenies. 
Am. Midi. Natural. 123:292-300. 

Williams. M.W., Jr. 1984. The cold hardiness adaptive response of green ash to geoclimatic gradients. 
Ph.U. thesis. The Pennsylvania State Universit>, 148 p. 

Wright, J.W. 1976. Introduction to Forest Genetics. Academic Press, New York. 463 p. 



128 



Steiner, K. C. and Carlson, J. E, eds. 2006. Restoration of American Chestnut To Forest Lands - Proceedings of a 
Conference and Workshop. May 4-6. 2004, The North Carohna Arboretum. Natural Resources Report 
NPS/NCR/CUE/NRR - 2006/001, National Park Service. Washington, DC. 

RATE OF RECOVERY OF THE AMERICAN CHESTNUT PHENOTYPE THROUGH 

_ BACKCROSS BREEDING OF HYBRID TREES 

% 

' "" Matthew B. Diskin and Kim C. Steiner 

School of Forest Resources, Pennsylvania State University, 
University Park, PA 16802 USA (steiner@psu.edu) 



Abstract: We describe a study in which the morphological characteristics of several chestnut populations 
- American chestnut, Chinese chestnut, first-generation hybrids, and first-, second-, and third-generation 
backcross hybrids - were quantified and compared. Twenty-four morphological variables known to 
distinguish American and Chinese chestnuts were used to develop a composite Index of Species Identity. 
The aggregate morphology of the first hybrid generation was almost exactly intemiediate (mean ISI = 
0.50) between Chinese (0. 1 1 ) and American chestnut (0.85). The first backcross generation resembled 
American chestnut more than expected, but the second and third backcross generations conformed closely 
to expectations. Although some degree of "Chinese" character could be found within the third backcross 
generation, 96 percent of the trees in this population fell within the range of ISI values for pure American 
chestnut and none fell within the range of Chinese chestnut. 

Keywords: Castanea dentata I backcross breeding / morphology / morphometries / Index of Species 
Identity 



INTRODUCTION 

Nearly a quarter of a century ago, Bumham ( 1981 ) proposed the backcross breeding system that is 
currently the basis for The American Chestnut Foundation's (TACF) chestnut restoration efforts. In a few 
years, TACF expects to produce a third intercross generation from the third generation of backcrosses to 
American chestnut (Hebard 2002). This BC3-F3 generation is expected to be highly blight resistant yet 
essentially "American" in all other characteristics, and it is expected that these trees will be the basis for 
the first serious efforts to restore American chestnut to its former habitats in the Appalachian region. 
Although experience with breeding other plants suggests that vhe third backcross is sufficient to recover 
the characteristics of the recurrent parent (American chestnut in this case), no one has yet quantified how 
well this will work in chestnut breeding. This is an important question for those whose interest is 
ecological restoration o^ American chestnut. How truly "American" will these trees be? 

In this paper we summarize a study that was designed to answer this question by comparing the 
morphological characteristics of American chestnut, Chinese chestnut, their first-generation hybrid (Fi), 
and three successive backcross generations to American chestnut (BCi, BC:, and BC3) (Diskin 2003). 
The morphology of the third backcross generation will be discussed in particular detail because this 
generation has the same relative proportion of the American chestnut genome as those trees that are 
currently proposed for use in restoration trials. 



OVERVIEW OF THE METHODOLOGY 

Twenty-four morphometric variables based on leaf, twig, bud, and stipule characteristics that distinguish 
American chestnut from Chinese chestnut were measured on trees sampled from TACF's Glenn C. Price 
Research Farm in Meadowview, Virginia. Approximately 50 trees, ranging in age from two to six years 



129 



old, were sampled from each of the following generations: American chestnut, Chinese chestnut, and F), 
BCi, BC2, and BC:, hybrid generations. All 24 variables were measured on each tree, and the results of 
the individual measurements were analyzed using standard statistical methods. 

The overall morphology of each tree was summarized in an "Index of Species Identity" (ISl). The ISI is 
the score of the first principal component, transformed to a scale from to 1 .0, from a principal 
components analysis of the 24 original variables. Essentially, ISI is a composite index of the best of the 
variables typically used by taxonomists to distinguish Castanca dentata (Marsh.) Borkh. from C. 
mollissima Blume. ISl score frequencies were plotted for each population, and the degree of o\erlap or 
separation in frequency distributions was used to compare the aggregate morphologies of the hybrid 
generations and their parental species. Mean ISI scores were also calculated for each generation and 
analyzed using standard statistical methods. 



MORPHOLOGY OF THE HYBRID GENERATIONS 
COMPARED TO THEIR PARENTAL SPECIES 

Because we measured only variables with proven utility in distinguishing Chinese and American chestnut 
specimens, the two species occupied the extremes of morphologies observed in the study. Chinese and 
American chestnuts scored at opposite ends of the scale for each individual \ariable as well as the 
composite variable, ISI (mean scores of 0.1 1 and 0.85 for Chinese and American chestnut, respectively). 

Comparative morphologies of the four hybrid generations and their parental species are most easily 
summarized by ISI scores. The morphology of the F| generation was almost exactly intermediate 
between American and Chinese chestnut, with a mean ISl of 0.50. The first-, second-, and third- 
generation backcross hybrids were different from the Fi hybrid but, surprisingly, similar to one another, 
with ISI means of 0.78, 0.77, and 0.79, respectively. 

Expected ISI scores can be calculated for each backcross generation assuming that observed ISl scores for 
the two species are accurate and assuming a straightforward 50 percent dilution of Chinese alleles in each 
backcross generation and quantitative, additive inheritance of ISl values. Under these assumptions, the F| 
should have an ISl that is exactly intermediate (0.48) between the parental species, and the obsen ed \alue 
of 0.50 is not significantly different from expectation. The BC|. BC:. and BC^ backcross hxbrids should 
have ISI means of 0.67, 0.76, and 0.81, respectively, or halfway toward the American species value of 
0.85 in each successive generation. 

Although ISI values for the BC^ and BC, populations were similar to one another. the\ did not differ 
significantly from the values expected under the above assumptions (0.77 \s. 0.76 and 0.79 vs. 0.81 for 
BCi and BC3 populations, respectively). The small difference between these two generations simply 
reflects the fact that backcrossing yields diminishing returns with each generation. However, the BC| 
population was anomalously more similar to pure American chestnut than expected (0.78 vs. 0.67). 
Among other things, the anomaly may be attributable to the fact that the 48 trees representing this 
generation were derived from crosses between only one Chinese and two American chestnut parents. Just 
as one or two individuals ma\ not be representative of an entire species, their progen\ ma\ not be 
representative of a typical hybrid population. (It should be noted that the h\brid populations used in this 
study are not the same as those used by TACF to produce its BC3F2 hybrids, nor are they directly related 
to one another in the sense, for example, that the particular BCi population in this study was used to 
produce the BC^ population that we measured.) 

Based on ISI values, 90 percent of the BC: trees and 96 percent of the BC3 trees had aggregate 
morphologies within the range of American chestnut values. None of the trees in these two populations 



130 



had the highest ISI vakies found in a very few American chestnut trees, and a small percentage had values 
lower than observed in any American chestnut trees. However, no backcross hybrid trees had values even 
close to the highest ISI values recorded for Chinese chestnut. 

American chestnut morphology was fully recovered in the BC3 generation for 15 of the 24 individual 
morphological characteristics that were measured. In each of these variables, there was no statistically 
significant difference between American chestnut and BC3 trees: leaf relative length, tooth length, tooth 
depth, leaf length to tooth length ratio, leaf width to tooth depth ratio, lenticel width, bud length, bud yaw 
angle, tooth hooking, leaf apex shape, interveinal leaf hairs, stipule size, twig color, twig hair density, and 
bud color. The BC3 generation did not fully resemble American chestnut in distance from base to 
maximum leaf width, twig diameter, bud width, bud relative length, bud appression, bud pitch angle, leaf 
base shape, leaf veinal hair density, and bud tip shape. However, for all but one of these variables (bud 
relative length), the third backcross generation more closely resembled American chestnut than Chinese 
chestnut. 



CONCLUSIONS 

Progress towards American chestnut morphology generally conformed to expectations based upon the 
proportion of American chestnut genome in the various hybrid generations: the F] was almost exactly 
intermediate between the parental species and the BC3 was very close to 15/16ths "American" on the 
composite index scale. Thus, backcross breeding appears to substantially recover American chestnut 
morphology in the backcross generations. Each of the three backcross generations was distinct from 
Chinese chestnut in that no individuals fell within the range of Chinese chestnut morphology, but each 
generation overlapped in morphology with American chestnut. Although the morphology of the third- 
generation backcross hybrids was largely similar to American chestnut, some Chinese-like characteristics 
remained. These could probably be further removed through selection for particularly "American" 
individuals in TACF's BC3-F2 generation before the production of BC3-F3 seed. 



ACKNOWLEDGEMENTS 

The authors sincerely thank the following people for their assistance on the project: Dave Armstrong, 
John Carlson, Fred Hebard, Benji Coniett, Peter Gould, and Amanda Subjin. We also thank the Schatz 
Center for Tree Molecular Genetics for financial support. 



LITERATURE CITED 

Bumham, C.R. 1981. Blight-resistant American chestnut: there's hope. Plant Dis. 65:459-460. 

Diskin, M. 2003. Morphological differences among Castanea dentata (Marshall) Borkhausen, Castanea 
moUissitna Blume, their first-generation hybrid, and three backcross generations. B.Sc. honors thesis, 
Pennsylvania State University, State College, PA. 47 p. 

Hebard, F. 2002. Meadowview notes 2001-2002. J. Am. Chestnut Found. 16(1):7-18. 



132 



sterner, is., l. ana Larison, j. t, eas. zvuo. Kestoration oi American LnestnuT lo rorest Lanas - rroceeaings or a 
Conference and Workshop. May 4-6, 2004, The North Carohna Arboretum. Natural Resources Report 
NPS/NCR/CUE/NRR - 2006/001, National Park Service. Washington, DC. 

SELECTION FOR CHINESE VS. AMERICAN GENETIC MATERIAL 

IN BLIGHT RESISTANT BACKCROSS PROGENY USING GENOMIC DNA 

. "^ Song Liu and John E. Carlson 

School of Forest Resources and Genetics Graduate Program, 
The Pennsylvania State University, University Park, PA 16802 USA (szil 10@psu.edu) 

Abstract: The American chestnut {Castanea dentata [Marshall] Borkhausen) was historically one of the 
most important hardwoods in North America due to its abundance and the multiple functions it served for 
both ecosystems and humans. The exotic chestnut blight fungus has eliminated the American chestnut as 
an overstory tree in eastern forest ecosystems, however. Backcross breeding shows promise to produce 
chestnuts that combine the blight-resistance that evolved in Chinese chestnut (C. moUissiiua Blume) with 
the desirable characteristics of American chestnut as a forest tree. In the backcross program, blight- 
resistance is introduced by an interspecific cross of American chestnut with resistant Chinese chestnut 
trees. American chestnut characteristics are then regained by a series of backcrosses to American 
chestnut parents. To accelerate and improve this selection process, we developed a molecular protocol to 
determine the amounts of American vs Chinese chestnut genome among progeny selected for blight 
resistance. The dot blot technique involves the hybridization of labeled Chinese chestnut genomic DNA 
to DNA from individual backcross progeny trees, which reveals the amount of Chinese chestnut DNA 
that remains in them. On average, progeny in the third backcross (BC3) generation should show lesser 
amounts of hybridization to Chinese chestnut genomic DNA probe than F2, BCl and BC2 progeny. 
Because there will be variation among individuals in each backcross generation for the amount of Chinese 
chestnut genome that they contain, those blight resistant progeny with greater amounts of Chinese 
chestnut content can be identified by this approach and eliminated from the crossing program. The 
effectiveness and reliability of this approach are demonstrated using samples from the parents and 
progeny in three backcross generations. 

Keywords: Backcross generations, dot-blot hybridization , genomic DNA, selection 

INTRODUCTION 

American Chestnut and the Chestnut Blight 

Before the introduction of the chestnut blight disease, the American chestnut (Castanea dentata 
[Marshall] Borkhausen) was one of the most important trees in hardwood forests of the eastern United 
States. With a range centered on the Appalachian Mountains and extending from Maine west to Michigan 
and south to Alabama and Mississippi (Little 1976), the American chestnut grew in mixtures with many 
other species, and often comprised 25 percent or more of the hardwood tree population within any given 
forest stand (Braun 1950). 

The American chestnut may have been the most important hardwood in eastern North America due to its 
abundance and the multiple functions it served for both ecosystems and humans (Hardin et al. 2001 ). It 
was a dominant component of much of the eastern hardwood forest, and it produced a regular and 
bountiful nut crop that was an important part of the diet of many animals (Rice et al. 1980). Historically, 
the American chestnut was an important tree because of the assortment of services and commodities it 
provided to people as well. It was an extraordinary tree for wood fiber production due to its large size, 
fast growth, and ability to sprout from stumps (Detwiler 1915). American chestnut wood fulfilled a 
multitude of needs ranging from construction and furniture lumber, firewood, fence construction, railroad 



133 



ties, telephone and telegraph poles, pulpwood, and tannins. The chestnuts were also important as a food 
source for rural residents, and the tree was widely planted to provide shade (Buttrick 1915). 

Chestnut blight was first introduced to North America in 1904. The chestnut blight disease is caused by 
Cryphonectria parasitica (MurnW) Ban i=Enc/othia parasitica [Murrill] P.J. and H.W. Anderson), an 
exotic fungus from Asia that enters through wounds in the bark and eventually girdles the tree, killing 
susceptible individuals (Roane et al. 1986). Because American chestnut trees evolved in the absence of 
the fungus, they lacked entirely any genetic protection from the fungus (Stiles and Hebard 1996). By 
1950 the disease had spread across the entire native range of the American chestnut, eliminating it as an 
overstory tree in eastern ecosystems (Newhouse 1990). The American chestnut continues to survive as a 
shrub, however, sprouting from the root collars of stumps in the forest (Hardin et al. 2001 ). ^_ 

Backcross Breeding Program 

The American Chestnut Foundation's (TACF) approach to developing the most resistant trees with the 
best American characteristics - ''the path of most resistance'" - is show n in Figure 1 . After the chestnut 
blight fungus was introduced to the United States, plant explorer Frank Meyer discovered the fungus in 
Asia, along with Chinese chestnuts (C. moUissima Blume) that had evolved resistance to the disease 
(Fairchild 1913). Because of the blight resistance of Chinese chestnut, and cold hardiness, this species 
was selected for developing blight-resistant hybrids with American chestnut that could replace the 
disappearing (Burnham 1987) in American forests. 




The Backcross Breeding Program 

(The Path of Most Resistance) 

^ ^ PARENTS 

CHINESc\ AMERICAN 

A Blight Susceptible 

Timber Tree 



Blight Resiiiant 
Orchard Tree 




Source of Resistance 



I Til Generation 




4 



Backcross Generations 



Second Generation 



BC1 



^^-^..^^ 



/ 

Third Generation DV^ i. '^ Crossed \Mth 

Fresh American 
Fourth Generation BC 3 Pollen 



Figure 1 . The scheme for the backcross breeding program being used b\ the .American 
Chestnut foundation (from Hebard. Http:// cheslnut.acf.org). 

However, while Chinese chestnut is highly resistant to the chestnut blight, it has other characteristics that 
make it undesirable as a replacement for the American chestnut. Whereas the American chestnut grows 



134 



straight and tall and was formerly a canopy tree species, the Chinese chestnut has a low-growing, 
sprawling form similar to that of an apple tree. Additionally, American chestnut trees have higher quality 
timber, sweeter nuts, and a faster growth rate (Hebard 1994a; Stiles and Hebard 1996). The genetic 
material of the American chestnut also reflects thousands of years of co-evolution with eastern hardwood 
forest ecosystems. During this time, complex relationships presumably evolved between the American 
chestnut and other components of the forest, a history that is borne in the genome of the American 
chestnut (Stiles and Hebard 1996). Thus a program based on back-cross breeding (Figure 1 ) was 
developed to recover the American characteristics while retaining the Chinese blight resistant genes. 

RATIONALE AND APPROACH 

The process of recovering the American characteristics by diluting out all of the Chinese donor parent 
characteristics, except for blight resistance, usually entails several generations of backcross breeding to 
recurrent parent trees (AC). The first hybrid generation (Fi) produced by crossing American chestnut 
with Chinese chestnut inherits one half of its genes from the American chestnut parent and one half from 
the Chinese parent. These first-generation hybrids are then backcrossed to an American chestnut parent, 
producing a first backcross generation (BCi) that has a genome that is on average three-quarters 
American chestnut and one-quarter Chinese chestnut. Each successive backcross reduces the Chinese 
fraction of the genome by one-half the second backcross generation (BCt) is on average one-eighth 
Chinese chestnut, and the third and final (in the plan outlined by Burnham) backcross generation (BC3) is 
on average fifteen-sixteenths American chestnut and one-sixteenth Chinese chestnut (Rutter and Bumham 
1982). However variation occurs among individuals in each backcross generation for the amount of 
Chinese chestnut genome that they contain due to chromosomal recombinations that naturally occur at 
gamete formation. In addition, TACF produces intercross (F2) generations (Figure 2) that increase the 
number of progeny at each generation, and provide greater genetic variation and greater opportunity for 
blight resistance to be separated from other tree characteristics. Selection for blight resistance and 



Intercross Generations 



f 



BC3 

Inoculate 
Sele:.-! 



EC 



Inoojiate 



Select 



Fifth Generation 



BC3F2 




PSU Arbordun-1 
PA-DC NR Penn Nursery 



Ino cu I ate 



;elei::t 



i 



Si>dh Generation BC 3 rJ 




f 



Seed Orchard 

PA-DCNR Penn Nursery 

Select / Certify/ Distribute 



f 



f 



T 

f 




Figure 2. Advanced generation intercross scheme of the American Chestnut Foundationn 
for seed orchard development and production (http:// chestnut. acf org) 



135 



tree characters is made at each breeding step, which requires intervals of several years. The breeding 
program could be accelerated through the use of genomics tools for the identification of trees carrsing 
larger portions of American genome at each step, thus also impro\ ing the results of each stage of 
selection. '" 

Many decades of breeding research by the U.S. Department of Agriculture, the Connecticut Agricultural 
Experiment Station (CAES), and the American Chestnut Foundation indicate that resistance in the 
Chinese species is carried on two or three genes, which are only incompleteK dominant. To achieve full 
resistance, all the genes from American chestnut that control response to the blight must be replaced by 
the Chinese alleles. The ACE breeding program has already reached the third backcross generation which 
is being evaluated in extensive field tests in several states for durability of resistance and for the 
American tall-timbered growth habit and regional adaptability. Overall. TACF has more than 1 1.000 trees 
at various stages of the blight resistance breeding process at its farms in Virginia. 

Random amplified polymorphic DNA (RAPD) and restriction fragment length polymorphism (RFLP) 
markers have been used to construct genetic linkage maps and identify genomic regions (QTLs) 
conditioning resistance in an F2 population derived from the "Mahogany" resistance source (Kubisiak et 
al. 1997). Two of theses regions have since been confirmed in a BCl population derived from "Nanking" 
and PI 34517 suggesting that some of the genomic regions conditioning resistance are s\ntenic across the 
different sources (Kubisiak, unpublished). AFLP markers were subsequently been found that fiank the 
blight resistance QTLs and that could thus also be used to select for those loci in progeny (Sisco, 
unpublished). 

However, while individual RAPD, RFLP, and AFLP markers will be good for early selection for the 
major resistance loci, such linked markers and associated maps with DNA markers will be difficult to use 
efficiently to select against the chromosomal material from Chinese chestnut that is not associated with 
resistance. To select against the Chinese genetic background, it will be necessary to use many markers 
covering all of the linkage maps simultaneously. When many markers are being used in concert, the 
inherent problems with reproducibility of RAPD markers, with dominance of the AFLP and RAPD 
markers, and with length of time and inconvenience needed to use RFLP markers would make selection 
against Chinese genetic background by the DNA marker approach ver\ complicated, and quite expensive. 

We have developed a simple dot blot protocol to rapidly screen individual trees in the breeding program 
for their content of American versus Chinese chestnut genome. The technique involves the h\bridization 
of labeled Chinese chestnut DNA to the DNA of individual trees, using American chestnut DNA to block 
the detection of sequences shared by American and Chinese chestnut. 

In the present study we tested the effectiveness and reliability of the dot blot technique to directly select 
against the Chinese genome in progeny of the BC3 generation. On average, progeny in the BC3 
generation should show lesser amounts of hybridization to the Chinese chestnut genomic DNA probe than 
F2, BC 1 and BC2 progeny. It should also be possible to identif\ those blight resistant progcn\ w ithin 
each backcross generation w ith greater amounts of Chinese chestnut content b\ this approach, so that they 
can be eliminated from the crossing program. 

The second objective of this study was to evaluate how closely the data from the dot blot protocol 
correlated with visual evaluation of known morphological characteristics. We assume that the \ariation 
of h\bridi/ation of American genomic DNA among individuals within each BC generation is coincident 
with the variation of the morphological characteristics that ta\onomicall\ distinguish American chestnut 
and Chinese chestnut. In the study of variation of the morphological characteristics among indiv iduals 
w ithin generations conducted by Matt Diskin (2003. and previous chapter in this proceedings). twent\- 
four morphometric characteristics known to discriminate between American and Chinese chestnut were 



136 



measured on each of approximately 50 individuals in the parental species, the first-generation hybrids, 
and in each of the three backcross generations. Principal components analysis was used to develop an 
Index of Species Identity (ISI) that described the aggregate morphology of the different populations. As 
expected, the morphologies of American and Chinese chestnut were the extremes measured in this study. 
The first-generation hybrids were intermediate between the two parental species, and the three backcross 
generations had similar morphologies, distinct from Chinese chestnut and largely similar to American 
chestnut. American chestnut morphology was essentially recovered in the third backcross generation, for 
the 24 characters studied. To find the relationship between variation in hybridization data and 
morphology, DNA was obtained from the sample individuals used in the morphology study. 

MATERIALS AND METHODS 

Chestnut Materials and Sample Selection 

Tissue samples were collected by Matthew Diskin (undergraduate thesis, PSU, December 2003) from 
trees at The American Chestnut Foundation's Glenn C. Price Research Farm in Meadowview, Virginia. 
Samples were taken from representative American and Chinese chestnut parents trees, their first- 
generation hybrids, and first, second, and third generation backcross hybrids (Table 1 .). 

Table 1. Populations sampled for morphology and dot-blot studies. 



Population 


Plantation and year 
planted' 


Years since 
planting 


Sample size for 
ISI study 


Sample size 
for dot blots 


American 


Amer2001 


2 


50 


10 


Chinese 


CbyCs 2000 


3 


49 


10 


F, 


MoreFls 1997 


6 


50 


10 


BC, 


JBls 1999 


4 


60 


30 


BC: 


JBls 1999 


4 


45 


26 


BC3 


Has 2000 


3 


49 


28 



'The plantation name refers to the chestnut plots at The American Chestnut Foundation's Glenn 
C. Price Research Farm in Meadowview, Virginia. 

The population of American chestnuts represented the open-pollinated progeny of seven chestnuts 
growing wild in Smyth County, Virginia. The population of Chinese chestnuts was composed of two 
unique pedigrees, derived from controlled pollinations between two different sets of Chinese parents. All 
American chestnut parents in the backcross generations were the plantation-grown progeny of open- 
pollinated trees growing wild in the mountains of Virginia, except that one was itself a tree growing wild. 
Neither the American nor Chinese chestnut parents that were sampled were used as parent trees in any of 
the hybrid crosses. Twelve pedigrees of first-generation hybrids were sampled. These trees were the 
progeny of nine Chinese chestnut mother trees and 12 American chestnut father trees. 

The populations of first-generation backcross trees sampled were progeny of a single American chestnut 
tree crossed with a single first-generation hybrid tree. Three pedigrees composed the population of 
second-generation backcross trees. The same first-generation backcross tree was used in each pedigree, 
but a different American chestnut parent was used in each cross. The population of third-generation 
backcross trees measured for this study comprised the progeny of a single second-generation backcross 
tree and a single American chestnut tree. There were no Chinese or American parents in common 



137 



between the first hybrid and any backcross generations or between the various backcross generations (see 
Hebard. this volume). ~ _ 

DNA Extraction, Digestion and Transfer ^^^"^ " 

DNA was extracted from twig samples that were selected for DNA dot blot analysis from among 10 
individuals among the chestnut parent and the first generation hybrid populations (Table 1 ). The samples 
were selected based on the Indices of Species Identity (ISI) determined by Diskin (2003) with the '- 

approximate ratio of 1 :2. Thus, the selected samples from the 3 BC generations should have the same 
distribution and population coverage as the original set of twig samples used by Diskin. The sample sizes 
in the 3 BC generations used for DNA extraction were: 30 samples in BCl. 26 samples in BC2, and 28 
samples in BC3. 

DNA extraction followed the manufacturer's instructions (Qiagen DNAeasy kit). Methods for DNA 
restriction enzyme digestion, agarose gel electrophoresis and alkaline transfer of DNA to nylon 
membranes were as described by Sharp et al. (1988). with minor modifications such as the use of Hybond 
N+ membranes (Amersham). Total genomic DNA was digested to completion using Hindlll restriction 
endonuclease (Gibco). The agarose gels were stained with ethidium bromide and onl\ those gels in which 
all tracks of genomic DNA showed approximatel) equal amounts of DNA after UV photography were 
used for transfer. 

DNA Quantification and DNA Dot Blot Preparation 

The individual tree DNA samples were quantified with a GeneQuant (Amersham) spectrophotometer 
(A26o)- All the DNA samples were diluted to 50ng/^L with ddH:©. Methods for manual preparation of the 
DNA dot blots followed the protocol provided by the nylon membrane manufacturer (Amersham). except 
that 1 |.iL of SOng/f^iL of denatured DNA sample was applied for each dot. in the simulation experiment, 
two repeated applications were applied to each dot, for a total of lOOng DNA. All the DNA samples were 
applied to the filters in a random order, following the random numbers generated by use of the MINITAB 
program (MINITAB 13.32, Minitab Inc. 2000). The applied ssDNA was fixed to the membranes using a 
UV crosslinker (Stratal inker®, Stratagene) for 30 sec. 

Probe Labeling and Southern Hybridization 

The labeling of probes with radioactive P-32, the h\bridization methods and the detection of 
hybridization signals followed manufacturer's instructions (Amersham). Bricflv. total genomic DNA was 
mechanically sheared by syringe, the length of probes was estimated by gel electrophoresis to be about 
500bp. The probes were denatured by boiling for 5 min and then labeled with P-32 b\ follow ing the 
random priming protocol (Invitrogen). The membrane was incubated at 65°C overnight in the 
prehybridization buffer with the denatured salmon spemi DNA. 

For experiments involving genomic blocking DNA, DNA fragments of 100-200 bp length were obtained 
by autoclaving the total genomic DNA for 2 min. The required amount of blocking DNA. 1-10 \\% niL" , 
was denatured by boiling for 10 min, added to the h\bridization buffer surrounding the membrane and 
incubated at 65°C overnight. The labeled probe ( 10-20 ng ml/') was added and the incubation continued 
for 8-16 hrat 65°C in the hybridization incubator. 

Washing and Signal Detection 

After hybridization, weakly hybridized and unhybridized probe was removed by three washes of 30 min 
each in 1) 2 X SSC (20 X SSC: 3M sodium chloride, 0.3 M sodium citrate. pH7)/0.l% SDS (sodium 
dodecyl sulphate) at room temperature; 2) 0. 2 X SSC/0.l%SDS at 42°C; 3) 0.1XSSC/0.I%SDS at 65°C. 



138 



Hybridization sites were detected using a phosphor imager after the membranes had been exposed to the 
imaging screen for 2 h. -_ 

Signal Normalization and Quantification 

Probe hybridization was measured quantitatively with a microcomputer-based image digitizing system 
TotalLab 2.00(Nonlinear Dynamics Ltd., 1996-2000). The intensity of the signals was digitized (Figure 
3), and each measurement was normalized to the values of the positive controls (Figure 4). To compare 
the digitalized signal data from each dot, nomialization was used to equalize the volumes in the dot 
images. This was accomplished by setting the nomialized volume of dots from a serial dilution to specific 
values (positive controls) and then recalculating all other volumes relative to those values. 




fgQQ 



E> 



-2' 

1. T«r«^ C^nrr* 



ki 



X 

Oror 



ti,rMm.-*nrri\e *1 A7. tl Cfl A5. « 



® 






Normalisation ^ 

■ I h- di^A:- v.. I.. 






IMJ fai^cWfndMv Mnc»ctlJ.^l(1jit 



miarH 



O (i) ® O © ® O O 

Q G GOO G O ® 

DO O G ® O O O 

^D O 



J 



iS 


Tnw7 




i.uirfi.#. 


'"* 


.,-fc.':«»Wl=«t'.»Javf;.'vi.'''-'---- '_■ ■■■ 




m 


a 








LMai 




T 


Vu^^ j 


vci-ek^#u j 


aac^<^^ 




*' 










520 *» 


sc£r4a 




ft 

M 

151 

TOO 
tSG 

114 


-J 


Al 












X&o 


In IT 


Ml 






AS 




a? 








•i-i 


^ c^,„ 


[■>:: 


fip««A Ajiif. 


-^ 'J 


• 













Figure 3. Screen capture of 
example signal quantification 
of dot blot using TotalLab 
2.00. 



I ^ L^ t^ ■•_) ^> 



isa^ 



X 






.■:MiP 


UOJit 


Normal lfiAt1<>n ^^ 




=™?~- 


NomitMbvlIni 




Clcicnttvpa 


',™r'° 



iimifws 



■ -ZZO. v-<^l. vaJ*-^9-C^<ti'flc«l -0 1X17 




Figure 4. Screen 
capture of signal 
normalization by 
TotalLab 2.00 for 
positive control 
dilution example. 



Statistical Analvsis 



13.32 



Statistical analysis of the normalized dot blot signal data was performed using Minitab version 
(Minitab Inc. 2000). Analysis of variance (ANOVA) was used to test the significance of mean 
differences within populations for signal intensity assuming equal variances. Brown and Forsythe's test 
was used to test for equal variance (Brown and Forsythe 1974). No data transformations were necessary. 



139 



Pearson's correlation coefficient was calculated to find the relationship between morphometric data and 
hybridization data (Minitab Inc. 2000). 

RESULTS ^=-^' ^ 

Differentiation of American Chestnut and Chinese Chestnut 

Preliminary experiments were conducted with parental DNAs to test the effectiveness of unlabeled 
American chestnut genomic DNA in blocking hybridization signal from shared sequences in the Chinese 
chestnut probe. The autoradiogram in Figure 5 shows the hybridization intensities obtained with labeled 
genomic Chinese chestnut probe hybridized to Southern blots of Hindlll digests of parental and backcross 
generation genomic DNAs after two lov\ stringencies washes. In the left panel of Figure 5. with no 
blocking DNA used, strong probe hybridization to DNA tracks from all generations is visible, and bands 
of restriction fragments from highly repeated DNA families are of similar intensity among samples. 
When the membrane was blocked with unlabelled DNA from American chestnut (right panel of Figure 5). 
the hybridization of Chinese chestnut DNA probe to the American and BC3 samples were greatly 
decreased, while the amount of hybridization to the Chinese. BCl and BC2 samples were decreased to a 
lesser extent. In addition, the intensity of hybridization for the smallest band in Figure 5 (arrovs ) is 
increased in the Fl and EC samples after blocking, while the American sample maintains the same low 
intensity, suggesting that this restriction fragment is Chinese -specific and when the American genomic 
DNA was blocked, the band was more accessible to the probe. 



5C1 BC2 BC3 C 



Fl EC1 3C2 BC3 




♦-I 



Figure 5. Southern Blot of genomic DNAs digested by Hindlll, and hvbridized against 
Chinese total DNA probe, labeled with P32. Left panel: No blocking DNA ; Right panel: 



American chestnut blocking DNA 



Signal Normalization With the Controls 



In this project, we normalized signals within blots by using a serial dilution of known amounts ol'Chinese 
chestnut and American chestnut DNAs on the blots as positive controls (figure 6). To avoid bias caused 
by experimental errors, the internal controls in each dot blot were used to nomialize the dot signals among 
blots probed by Chinese total DNA. with American blocking DNA. 



140 




Figure 6. Autoradiogram of different amounts of Chinese DNA vs. Chinese total DNA 
probe, using American blocking DNA. This is used as a positive control to normalize dot 
signals. 

Hybridization Variation Within and Between Chestnut Generations 

A simulation experiment was conducted to test the level of sensitivity of the dot blot technique to genome 
variation among the chestnut generations. In the simulation experiment, mi.xtures of Chinese and 
American total DNAs equal to the average expected ratios for the Fl and 3 backcross generations (1:1 for 
F 1 ; 1:3 for BC 1 ; 1 : 7 for BC2; 1:15 for BC3 ) were used to simulate average genome content in each 
generation. On the same blot, an equal amount of genomic DNA pooled from 5 individuals from each 
generation was applied and probed by Chinese total DNA with blocking DNA from American chestnut 
(Figure 7). As expected, the American chestnut DNA dot has the least signal intensity, while the Chinese 
DNA dot has the strongest signal. From Fl to BC3, the intensity of the dots decreased proportionately. 
When the amount of hybridization to the dots from the simulated DNA admi.xtures and the bulked DNAs 
were compared using TotalLab image analysis software, the results showed that the real and simulated 
mixtures had the same levels of intensity (Table 2), suggesting that on average, the backcross generations 
have the same ratio of Chinese chestnut genome and American chestnut genome as expected, which the 
dot-blot technique can faithfully detect. 

To detemiine the extent of variation among individuals within and among backcross generations, a new 
blot was prepared with DNA dots from 10 individuals from BCl, 12 individuals from BC2 and 12 
individuals from BC3, plus the parental DNAs as the internal controls. This blot was probed with labeled 
Chinese chestnut total DNA, blocked with unlabeled American Chestnut DNA (Figure 8). The signal 
intensity of each dot for this hybridization was measured in TotalLab (Table 3), and compared to the 
internal controls. The relative signal intensity, following TotalLab normalization, measured 328 for 
Chinese DNA and 19 for the American parental DNA. From the distribution of signal intensities among 
generations (Figure 9), we found, on average, that the BCl individuals have stronger hybridization than 
BC2, while BC2 have stronger hybridization than BC3. This trend is as expected from the backcross 
program, i.e. that in general BC3 individuals have the least Chinese genome DNA remaining. However, 
much variation in hybridization was detected within each generation, opening the possibility for selection 
based on DNA content. 



141 



c 


A 


1:1 


1:5 


1:" 


1:15 


t 


• 


• 


• 


• 


• 


• 


• 


i 




t 


• t 


c 


A 


Fl 




BCl 


BC2 BC3 



Figure 7. Autoradiogram of Dot 
Biol Inbridization of mixtures of 
Chinese and American total DNAs 
vs. Chinese DNA probe, with 
blocking DNA from American 
chestnut. First Line: Mixtures of 
DNAs in the average expected 
ratios for F I (1:1) and the 3 
backcross generations (1 :3 for 
BCl; l:7forBC2: 1:15 for BC3). 
Second Line: Bulked DNA 
samples of six indi\ iduals from 
each generation. Each dot had 100 
ng of genomic DNA delivered in 
2uL. C. Chinese: A, American; 
Fl. F I generation: BCl, 
Backcross I generation: BC2, 
Backcross2 generation; BC3, 
Backcross3 generation. 









- 


- 




- 






^ 




*■>• 


j^ 


t 


f 




.i^f 


# 



Figure 8. Autoradiogram ol Dot Blot indi\idual DNAs from BCl (10 samples). BC2 (12 samples) 
and BC3 (12samples) vs. Chinese total DNA probe with blocking DNA from American. Lach dot 
has 50 ng of genomic DNA delivered in I uL, all the samples are randomly arranged in this array. 

To determine if the differences in hybridization intensities were statistically significant among the 
backcross generations, we used ANOVA in MINI! AB to analyze the measurements b\ generations. For 
the result of one-way ANOVA (Figure 10). the P-value was 0.0000. showing that the \ariation of 
hybridization among the generations was highly significant. In the dotplot graph of the h\bridization 
measurements (shown in Figure 1 1 ), the mean of the BCl values was significantK greater than BC2, and 
BC2 was only slightly greater than BC3. For the ANOVA analysis, the variation within each generation 
was assumed to be equal. The statistical test for equal variation showed that the variation was indeed 
equal in each generation, although there were two individuals with greater \ariation than others in BCl, 
suggesting that there were some experimental errors or random errors in the procedure in tiiose cases. 



142 



When we increased the sample size (Figure 13), however, the random errors were much smaller than in 
the experiment with smaller sample size. 




DAdmixtures 
■ Experiments 



Table 2. 
Histogram of 
nomialized data 
for admixtures 
and 

experimental 
samples in 
Figure 7. 





generations 


BC1 


BC2 


BC3 


1 


156 


94 


70 


2 


121 


1 10 


38 


3 


192 


79 


33 


4 


142 


104 


31 


5 


148 


118 


39 


6 


294 


81 


32 


7 


140 


119 


28 


8 


286 


52 


18 


9 


163 


111 


57 


10 


159 


65 


36 


11 




68 


92 


12 




79 


47 



Table 3. Values for 
nomialized signal intensity 
data of hybridization shown 
in Figure 7 (Control values: 
19 for American, 328 for 
Chinese DNA). 



350 n 

300 
250 




































■ 




■ 










■ BC1 


200 


" 


■ 




















A BC2 
BC3 


150 ' 
100 


1 
■ 


▲ 


■ 
A 




▲ 


■ 
A 




■ 
A 


■ 




i 


^^"American 
^^"Chinese 


50 - 


- 












A 




A 






■ 







t 


' 


1 


1 


1 


1 


1 


1 


; " 




1 2 


3 


4 


5 


6 


7 


8 


9 


10 


11 


1 


2 



Figure 9. Signal 
discrimination for dot 
blot ii.tensities between 
and within each BC 
generation, with 
American and Chinese 
parent as controls. 



143 



One-way 


ANOVA: nomalizations versus generation 






Analvsis 


jf Variance fcr 


r.rraaliza 








Scurce 


DF 53 


1-lS 


F P 






generari 


2 103780 


51850 


37.04 0.000 






Error 


31 43430 


14G1 








rctal 


33 147210 




Individual 5»5> CIs ::i 


Mean 






N Mean 


StDev 


Based ca Pooled Stjer 














1 


10 130.10 


€0.76 




(- 


) 


A 


12 90.00 


22.47 


( i 






3 

Pcolsd St 


12 43.42 

Z^v = 3". 43 


20.55 


f- * — ) 






5C 100 


150 


200 



Figure 10. One-way ANOVA test shows that differences in the signal intensity data 
between generations are significant. 



300- 




(group means are indicated by lines) 






8 








c 200- 


o 








nomalizat 

o 

1 


E 




i 



o 


o 
o 

1 




0- 

generation 






5 




1 


1 

CM 


1 

CO 





Figure 1 1 . Dotplot of normalized dot blot data by BC generation, 
mean values tor each BC generation. 



Red bars indicate 



Relationship Between Variation in Hybridization Signal Intensities and Morphological Variation Amoim 
Backcross Generations 

An inherent assumption with use of the dot blot protocol to screen for individuals with greater amount of 
American chestnut DNA w ithin the backcross generations, was that a strong relationship should exist 
between variation in DNA at the genome level and the phenotypic. or the morphological, variation within 
BC generations. Diskin et al. (2006) (and pre\ ious chapter) measured t\\ent>-four discriminating 
morphometric characteristics in each of the parental species, the firsl-generation li\brids. and the three 
backcross generations. Diskin used principal components analysis to develop an Index of Species 



144 



<Z- l-» i r-»^ S'^S' ^|->^^t»-»eit 



LL 



I — ly l^i'icl < 



30 



'1=rFr 



Sd , I— lyl:>ricls. 



«=> 



" 


















































r-r 


1 — 1 . , i 



>C;= I— lyl=.i-icl! 



cl ) 3S 































— 














1 




H t- 


1 



I — ly l3 ric:l * 



^ > 3S 

30 



xx 



,.^\, m ^r I" i d: ,3» n c_: l-ns-s-tn »_it 



<^> 





1 








1 








































-r 


~H 


o - 


_ — . — , — , — , — , , — , . — , , — . . — . — . — .III 



<=»r CI>|=>€r^l^^ (cl^ritlty 



Figure 12. Index of 
Species Identity. 
Frequencies of the 
Index of Species 
Identity scores are 
plotted along the y- 
axis (from Diskin's 
thesis). 



Identity (ISI) that described the aggregate morphology of the different populations relative to American 
and Chinese chestnut phenotypes. As expected, the morphologies of American and Chinese chestnut 
were at the extremes measured in his study. The first-generation hybrids were intermediate between the 
two parental species, and the three backcross generations had similar morphologies, distinct from Chinese 
chestnut and largely similar to American chestnut. American chestnut morphology was essentially 
recovered in the third backcross generation, based on ISI. To determine the relationship between variation 
in the genomic DNA hybridization data and the morphological variation, we prepared a DNA dot blot 
with 2 or 3 trees sampled from each bin of the frequencies of the ISI in each BC generation (Figure 12). 
The hybridization result is shown in Figure 13. All the samples were arranged in random on the blot, with 
three replications to decrease the hybridization bias and experimental error. The Pearson's correlation 
coefficient was calculated between the normalized signal intensity measurements and morphological ISI 
for each individual tree, yielding a value of -0.662, which is statistically significant (Figure 14). 



145 




Figure 13. Dot blot of selected samples (82 individuals with 8 controls) from the Diskin 
morphometric study (Complete random design with two replicates) probed with blocked, 
total Chinese DNA. 















350 — 




* 








300 - 












(fl 250 — 












c 




♦ ♦♦ 








o 




4 








cS 200 - 




•♦ * > 








P 150 — 
O 




♦ ♦ ♦ 










**.* ♦♦. * ♦ * ♦ ♦ 








^ 100 — 




♦ ♦♦'■♦ 

♦ ^ ♦ .**♦♦ 

♦ / ♦♦ ♦ ^ • • 

# ♦ ♦ ♦♦ 




♦ 




50 — 




* ♦♦ ♦ ♦ 


• 








♦♦ / «• V 


* 


♦ ♦ 








* 








— 












1 
0.2 


1 1 1 1 

0.3 4 5 6 

index 


1 

07 


1 

08 



figure 14. Plot of signal intcnsit\ data vs. morphological index data. 
(Pearson correlation of normalizations and index -0.662: P-Value = 0.000) 



146 



^— ^ DISCUSSION AND CONCLUSIONS 

Genome dot Hybridization Protocol 

Genomic hybridization involves extraction of genomic DNA from one of the species of interest, for use as 
a probe by either Southern hybridization to DNA blots or by in situ hybridization to chromosome 
preparations from the species or hybrids being studied (Orgaard and Heslop-Harrison, 1994). Many of the 
DNA sequences within the two genomes under investigation may be sufficiently different so that genomic 
probing discriminates them. Those differences between species may include the different members of 
classes of repetitive DNA and species-specific DNAs. We do not know the detailed genomic differences 
between Chinese chestnut and American chestnut, but the results from the hybridization to DNA digests 
without the blocking DNA showed that their genomes have a high level of similarity, as expected for 
closely related species. The addition of an excess of unlabelled DNA from the American chestnut parents 
(blocking DNA) in our experiments substantially increased the specificity of the probe, enabling the two 
species to be distinguished by hybridization to DNA digests or dot blots. The effect of blocking in our 
experiments may be due to (a) hybridization between probe DNA (Chinese chestnut DNA) and common 
sequences in the blocking DNA (American chestnut DNA), (b) hybridization between the blocking DNA 
(American chestnut DNA) and common sequences on the membrane-immobilized DNA (Dot blot DNA) 
or (c) a combination of both. 

The use of total genomic DNA, in combination with blocking, as a species-specific probe has several 
advantages. The use of genomic DNA as a probe avoids the need for the time-consuming and uncertain 
process of screening DNA clones from a library for clones that are specific to the American or Chinese 
chestnut genomes. Furthermore, it would not have been possible or practical to find enough American or 
Chinese specific sequences to cover those genomes in the present study. In contrast, the use of genomic 
probes is simple and straightforward in application, making it practical to develop a screening piotocol for 
application within a large backcross breeding program. 

Hybridization Variation Within and Among Backcross Generations 

Because the backcrosses were made only to American parents, the Chinese chestnut genome was 
expected to be progressively diluted as backcrossing progressed. Statistically, it was expected that 
American genome should comprise on average half of the genome of individuals in the first interspecific 
hybrid, three-fourths of the genome of the first hybrid generation backcrossed to American, seven-eights 
of the genome of the first backcross generation backcrossed again to American, and fifteen-sixteenths of 
the third backcross population, if we assume the parents species have totally different genomes. Also, the 
variation of genome amounts should become smaller and smaller within each BC generation following 
successive selections for blight resistance and tree phenotypes. Correspondingly the difference in level of 
hybridization signal on dot blots should also be observed to decrease in magnitude between generations 
from the BCl to BC3 generations (1/4 ^ 1/8 -> 1/16). 

Like Diskin's phenotypic ISI index, the DNA dot blot results with American and Chinese chestnut 
parental species trees in this study were distinct, and represent the two extreme cases, as shown by their 
scores in Figure 9. This is reasonable, and expected, as the study was based on the known genome 
differences between American and Chinese chestnut. 

The hybridization signals of the populations measured in this study were summarized by their nomialized 
data (Table 3). The progression towards American-like genome in each successive hybrid generation from 
BCl to BC3 was apparent from the decrease in the means of the normalized signals (Figure 1 1 ). Also, as 
expected, the decrease in mean values from BCl to BC2 was much greater than the decrease in values 
from BC2 to BC3. The decrease in magnitude of change towards the American chestnut genome value 



147 



among backcross generations fits expectations: each successive backcross generation is on average more 
American than the previous generations and the genome of the third backcross generation (mean = 43.42) 
approaches most closely that of American chestnut (normanzation=19). 

The ANOVA resuhs showed that the mean differences among the BC generations were significant. This 
proved that the variation in dot blot hybridization is related to genomic variation, and not caused by 
experimental errors or random errors. The ANOVA test result of equal variance among each BC 
generation is not what we expected based on statistical considerations, however. The reasons for the equal 
variance in genome content among generations may be ( 1 ) that the sample size was not big enough to 
represent the whole population, bringing bias into the population sampling: or that (2) the genome 
differences between Chinese and American chestnut are actually too small to reliably distinguish the 
variances among the generations at the scale of dot blot sensitivit> or that (3) additional variation is 
produced at each generation by recombination events during gamete formation. 

Relationship Between the Hybridization Data and Morphology Data 

To be able to screen for the individuals which are more American- like in the BC generations based on 
the results of dot blots, one should show that the variation in DNA content within generations is strongly 
related to the morphological variation. The ideal result would be a one to one relationship (Pearson's 
correlation coeftlcient=l or - 1 ). 

In the project conducted by Matthew Diskin (Diskin 2003), an Index of Species Identit\ (IS!) was used to 
describe the aggregate morphology of the different populations. In our study prepared dot blots from 
samples selected from among those used by Diskin. The relationship between the dot blot h> bridization 
data that we obtained and Diskin's ISl values was strong (Pearson's correlation coefficient = -0.662). A 
possible reason for this strong correlation could be that most of the genome sequences detected b\ the dot 
blot technique are expressed coding sequences that evolved at the same rate or along with the evolution of 
the morphological differences between the species. The negative \alue of the relationship is logical, 
because the ISl is positively related to American characteristics, while the hybridization data is negatively 
related to the amount of American genome DNA. 

The relationship between morphological index and genomic dot blot signal intensities was not one to one, 
however, indicating that it is not possible to predict the morphometric differences between trees w ith 
100% success based just on the differences in dot blot signals. Two possibilities could account for this. 
The first possible explanation arises from the fact that not all of the morphological \ariation that 
represents the species-specific characters were used to generate the ISl. If the morphological 
characteristics measured were not comprehensive enough, this could cause a bias in the ISl analysis. The 
genome-level variation assessed by dot blots should, in theorv, be able to uncover differences in many 
more characteristics than is possible though phenoty pic evaluation. 

A second possible explanation for the differences between the ISl and dot blot results could be that the 
genomes of Chinese chestnut and American chestnut are highly similar because they are closely related. 
DNA sequences in the Chinese chestnut genomic probe that are highl> similar to American chestnut 
sequences will be removed during the blocking step even though they may have ver>' different expression 
patterns and cause different morphological characteristics, fo minimi/e this concern, high stringencies 
were used in the dot blot filler washing steps and in probe blocking to ensure that onl\ \ irtuall\ identical 
sequences between American and Chinese chestnut species would be removed from the genomic probe. 

In summar>\ this project has demonstrated that the dot blot technique can produce similar results to that 
obtained b\ the more painstaking and lengthy assessment of genotypes based on assessment of 
morpholog) for individuals in American chestnut backcross generations, fhc convenience, sensitivit\ 



148 



and rapidity of the dot blot approach should make the technique more suitable than phenotyping for 
screening large populations of trees and seedlings for American vs. Chinese genetic makeup. The 
observation that a significant amount of variation in dot blot signal intensity was observed among 
individuals in all three of backcross generations, indicates that the dot blot technique would be useful for 
selecting individuals with the greatest amount of American genome at each generation. The dot blot tool 
could thus greatly accelerate the goal of breeding blight resistant trees that have regained the genetic 
makeup of the American chestnut species. 



ACKNOWLEDGEMENTS 

We gratefully acknowledge the financial support of The American Chestnut Foundation and the Schatz 
Center for Tree Molecular Genetics, the provision of samples by Fred Hebard, and the technical 
assistance of Matt Diskin and Mike Porzio. 



LITERATURE CITED 

Anamthawat-Jonsson, K., and J.S. Heslop-Harrison. 1992. Species specific DNA sequences in the 
Triticeae. Heriditas 1 16:49-54. 

Anamthawat-Jonsson, K., T. Schwarzacher, A.R. Leitch, M.D. Bennett, and J.S. Heslop-Harrison. 1990. 
Theor. Appl. Gen. 79:721-728. 

Braun, L.E. 1950. Deciduous forests of eastern North America. Blakiston Company, Philadelphia, PA. 

Brown, M.B., and A.B. Forsythe. 1974. Robust tests for equality of variances. J. Am. Stat. Assoc. 69:364- 
367. 

Bumham, C.R. 1987. Historical overview of chestnut breeding in the United States. J. Am. Chestnut 
Found. 2(1 ):9-ll. 

Buttrick, P.L. 1915. Commercial uses of chestnut. Am. Forestry 21(262):960-968. 

Detwiler, S.B. 1915. The American chestnut tree. Am. Forestry 2l(262):957-960. 

Diskin, M. 2003. Morphological differences among Costanea dentata (MarshaW) Borkhausen, Castanea 
mol/issinia Blume, their first-generation hybrid, and three backcross generations. B.Sc. honors thesis, 
Pennsylvania State University, State College, PA. 47 p. 

Diskin, M., K. C. Steiner, and F. V. Hebard. 2006. Recovery of American chestnut characteristics 
following hybridization and backcross breeding to restore blight-ravaged Castanea dentata. For. Ecol. 
Manage. 223:439-447. 

Hardin. J.W.. D.J. Leopold, and F.M. White. 2001. Textbook of dendrology. McGraw Hill, New York. 

Hebard, F. 1994. The American Chestnut Foundation breeding plan: beginning and intennediate steps. J. 
Am. Chestnut Found. 8( 1 ):2 1 -27. 

Hebard, F. 2003. Differences between Chinese and American chestnut. Unpublished manuscript received 
through personal communication. 



149 



Http:// chestnut.acf.org. 2004. The American Chestnut Foundation. Accessed April 1 5, 2004. 

Kubisiaic, T.L., F.V. Hebard, CD. Nelson, J. Zhang, R. Bernatzky, H. Huang, S.L. Anagnostakis, and 
R.L. Doudrick. 1997. Molecular mapping of resistance to blight in an interspecific cross in the genus 
Castanea. Phytopathology 87(7):75 1-759. 

Little. E.E. 1976. Atlas of United States trees: Vol. 4, Minor eastern hardwoods. USDA. Misc. Pul. 1342. 
United States Government Printing Office, Washington D.C. 

Orgaard, M., and J.S. Heslop-Harrison. 1994. Relationships between species of Leymus, 
Psathyroslachys, and Hordeum (Poaceae. Triticeae) inferred from Southern hybridization of genom.ic and 
cloned DNA probes. Plant System. Evol. 189:217-231 

Newhouse, J.R. 1990. Chestnut blight. Sci. Am. 263(1): 106-1 11. 

Rice, G., A. McCoy, T. Webb, C. Bond, and V. Speed. 1980. Memories of American chestnut. P. 397-421 
in Fo.xfire 6, Wigginton, E. (ed.). Anchor Press/Doubleday. Garden City, NY. 

Roane. M.K.. G.J. Griffin, and J.R. Elkins. 1986. Chestnut blight, other Endothia diseases, and the genus 
Emluthia. APS Press, St. Paul, MN. 

Sharp. P. J.. M. Kreis, P.R. Shewry, and M.D. Gale. 1988. Location of B-amylase sequences in wheat and 
its relatives. Theor. Appl. Gen. 75:286-290. 



150 



Steiner, K. C. and Carlson, J. E, eds. 2006. Restoration of American Chestnut To Forest Lands - Proceedings of a 
Conference and Workshop. May 4-6, 2004, The North Carohna Arboretum. Natural Resources Report 
NPS/NCR/CUE/NRR - 2006/001, National Park Service. Washington, DC. 

BIOLOGICAL DIMENSIONS OF THE GMO ISSUE 

_:: John E. Carlson 

,_ ^ School of Forest Resources, The Pennsylvania State University, 

University Park. PA 16802 USA (jecl6@psu.edu) 



Abstract: Genetic engineering is opening new opportunities in tree improvement, including access to 
novel genes for disease resistance. American chestnut is one of the tree species for which genetic 
transformation has been reported. Thus restoration of American chestnut could benefit from the genetic 
engineering technology, should suitable genes for blight resistance be identified. However, questions 
regarding the efficacy and safety of genetic engineering with trees, and other perennial plant species, have 
been raised including the stability of integration and expression of foreign genes, long term effects on 
non-target species, transgene dispersal through seed and pollen, etc. Many of these concerns could be 
alleviated if reliable approaches were available for the confinement of transgenes. The National Research 
Council assembled a committee of twelve researchers from various disciplines to evaluate the measures 
for biological confinement of transgenes. This paper will summarize the committee's findings regarding 
the unique concerns that arise in the application of GE to trees, and the potential that exists for applying 
bioconfinement methods with GE trees. 

Keywords: National Academies; genetic engineering; trees; biological confinement; transgenes 



INTRODUCTION 

The ability to apply GE to plant improvement extends beyond annual crops to woody perennia' plants 
including forest trees (reviewed by Pena and Seguin 2001 ) and fruit and nut trees (reviewed by Trifonova 
and Atanassov 1996) such as American chestnut (Conners et al, 2002a.b). E.xamples of successful GE 
with trees are appearing with increasing frequency in the scientific literature. Applications for field trials 
of GE trees have been submitted in the US, Canada, and the EU over the past 10 years for a wide variety 
of tree species ranging from pines to persimmons, and poplar to papaya (www.isb.vt.edu/cfdocs/ 
fieldtestsl.cfm). 

The Committee on the Biological Confinement of Genetically Engineered Organisms was established by 
the National Research Council in 2002 to evaluate the status of and potential role of biological 
confinement technologies for Genetically Engineered Organisms (GEOs). The study was solicited by the 
U.S. Department of Agriculture. The National Research Council assembled a committee of 12 e.xperts 
from various disciplines. The committee prepared its report over a 15 month period, meeting several 
times at the facilities of the National academies if Science in Washington DC and Irvine, California. In 
2004, the Committee on the Biological Confinement of Genetically Engineered Organisms published its 
findings (Kirk, Carlson, et al 2004). This report is available in hardback edition from the National 
Academies Press, or as downloadable files at the NRC web site (http://www.nap.edu/books/0309090857/ 
html/). This paper will review the background rationale for the study, the findings of the committee 
regarding trees and other woody perennials, and the committee's recommendations. 

The federal policy framework for the regulation of biotechnology products was created in 1986. This 
framework followed a vigorous discussion amongst scientists and the public over a period of several 
years, and has continued to be debated since. Many new GEOs have been developed since the framework 
was established. More recently the application of GE technology has expanded from annual crops to 



151 



perennial plants and animals, including forest trees such as American chestnut. The expansion of GE to 
long lived organisms raised the issue amongst the regulator) communitv of how best to ensure 
confmemcnl of transgcnes over the long term (addressed in McLean and Charest 2000). Examples of 
failure of physical confinement protocols and the concerns raised in the public over the "escape" of 
transgenes into the food supply and the environment, suggested that it was time to also look at biological 
confinement opportunities. 

The committee was charged with addressing the following questions: 

1 ) What bioconfinement methods for GEOs are available, and how feasible, effective and costlv are 



these methods' 



2) What do we know about when and why methods fail, and what can be done to mitigate those 
failures? 

3) When these methods are used in large-scale applications. v\hat procedures can be used to detect 
and cull individuals for which the bioconfinement methods have failed? 

4) What are the probable ecological consequences of large-scale use of bioconfinement methods at 
the population, community, and landscape levels? 

5) What new data and knowledge are required for addressing these important questions? 

6) What is the social acceptability of bioconfinement methods? 

The committee focused on risks associated with the dispersal of a transgene or transgenic organism into a 
place, population, or biological community for which it was not intended. Long-lived species that 
disperse easily can present particular risks due to the inefficacy of physical confinement methods, and the 
potential for escapees to interact with wild populations. Thus the committee was asked to pay particular 
attention to transgenic fish, shellfish, trees, and grasses. 

The concerns most often associated with risk of dispersal of transgenes include 1 ) The evolution of 
increased weediness or invasiveness; 2) Effects on nontarget populations — including humans; and 3) 
The potential for transgenes to disperse into the environment during field tests before being deregulated. 
The impact of transgene dispersal from perennial plants such as trees often involves issues of gene fiovv 
into natural populations. Gene fiow issues with GEOs have been addressed in several studies, including 
trees (Slavov et al. 2002) from both theoretical and applied perspectives, and will not be dealt with much 
here. 



WHEN WILL BIOCONFINEMENT BE NECESSARY FOR TREES? 

GE trees are appearing with increasing frequency. Since 1989. more than 230 pemiits for field tests of 
GEO plants have been approved by the Animal and Plant Health Inspection Ser\ ice in the United States, 
including 1 8 woody plant species. At least 65 permits have been granted in other countries for field trials 
on GE trees and other woody plants. The tree species for which field tests have been approved in include 
pines, persimmons, apples, walnuts, spruces, Sweetgum, aspen, plum, poplar, pear, and papaya 
(wwvv.isb.vt.edu/cfdocs/fieldtestsl.cfm). 

Numerous traits are being engineered in trees, including lignin modification, increased growth and 
productivity, enhanced utilization of resources, pest and disease resistance, stress tolerance, herbicide 
resistance, optimization of mycorrhizal symbioses, phytoremediation of contaminated soils, and even 
production of anticancer drugs (Sederoff 1999; Merkle and Dean 2000). 

Specific concerns arise regarding the potential for escape of transgenes w ith forest trees. One concern is 
that long-distance gene movement occurs naturalK with trees. Pollen and seeds from trees can be carried 
very long distances by wind, animal, and water vectors. Also, interfertile wild or feral relatives are quite 



152 



common among tree species. Since hybridization can occur so readily among trees, even exotic trees 
used for GE may require confinement if within pollen flow distance to related tree species. Another 
important issue is that trees are more likely to be keystone species within their ecosystems than other 
plants and animals. When keystone species are impacted, invasion and non-target impact consequences 
can be large and spread far beyond the species itself Finally, forests trees carry a higher level of societal 
importance and impact than most agricultural species. Concerns over forest health are more than 
economic; they extend to issues of importance of place, esthetics, recreation, nature, and ecosystem 
services. 



BIOCONFINEMENT TECHNOLOGIES 

The committee searched the literature to identify as many bioconfinement methods as possible. Each 
technique or potential technique was described and its strengths and weaknesses evaluated. The 
following bioconfinement methods were reviewed: 

Sterility 

Mortality of Vegetative Propagules 

Confining Pollen-Mediated Spread of Transgenes 

Transgenes Absent from Seeds and Pollen 

Artificially Induced Transgene Expression 

Reducing Gene Flow to Crop Relatives 

Repressible Seed Lethal Confinement 

Cross-Incompatibility 

Fitness Reduction in Transgenic Crop-Wild Progeny 

Phenotypic and Fitness Handicaps 

Reduced Exposure to Transgenic Traits 

The category of sterility, included the use of interspecific hybrids, sterile triploids, unisexual plants 
lacking mates, transgenic sterility, ablation of reproductive organs, and reversible transgenic sterility (a 
"GURT" or Genetic Use Restriction Technology). 

A GURT approach to bioconfinement of transgenes that has received a great deal of attention and 
research is known as "trait-genetic use restriction technology.'" Trait- GURTs are alternative approaches 
to reduce the effects caused by unwanted transgenes by activating a transgenic trait at a specific time 
through a specific artificial stimulus, such as a chemical spray. In this way, non-target organisms are 
spared long term exposure to the trait, the targeted species may be less likely to develop resistance, and 
GEOs that escape confinement should not have any selective advantage as the trait would not be 
expressed in the absence of the inducing agent in nature. 

Several opportunities for confining pollen-mediated spread of transgenes were considered such as 
nontransgenic male sterility, transgenic male sterility, transgenes in chloroplast DNA transgenes in 
chloroplast DNA, and apomixis (for asexually produced seeds). 

Creating GEOs in which the transgenes are absent from seeds and pollen has been proposed by use of 
non-transgenic scions on transgenic rootstock, and by programmed excision of transgenes before 
reproduction. Also, programmed cell death was evaluated as a means to induce mortality in vegetative 
propagules. 

Approaches to reduce exposure of non-target organisms to transgenic traits through tissue-specific gene 
expression have also been proposed, including chloroplast-targeting of gene expression, limiting 



153 



expression to roots and tubers, vascular tissue-specific gene expression, flov\er- and fruit-specific gene 
expression, pollen-specific gene expression, and seed-specific gene expression. 



EXAMPLES OF BIOCONFINEMENT METHODS 



Sexual reproduction of genetically engineered plants can be blocked by including a gene that renders the 
organism either permanently sterile (nonreversible transgenic sterilitv ) or conditionally sterile until an . 
appropriate trigger is applied, such as the use of a chemical sprav on a plant (reversible transgenic 
sterility). 

Engineered sterility has its strengths and weaknesses for bioconfinement of GEOs. The main strength of 
the approach is in its overall effectiveness of confinement. Since with trees the primary risk of transgene 
escape is through pollen and seed, reliable sterility could overcome much of the risk associated with GE 
trees (Strauss, et al. 1995). However, there were weaknesses found in the general application of sterility 
as well, such as concerns that engineered sterility has not been adequately tested yet to determine how 
effective and reliable it will be for long-lived organisms, that it may be unsuitable for farmers who save 
seed from specific crops for replanting the next year, and it is still uncertain how well engineered sterility 
will be accepted by the public. With trees, the specific concern would be those cases in which the seed or 
fruit crop may be a highly desirable feature, such as in the production of mast for wildlife, in which cases 
engineered sterility would be counterproductive, unless reversible. 

Although most of the above approaches are still untested in the field, their efficacy is generally 
recognized as having great potential. The options for engineering the nuclear genome to effect sterilit> in 
trees that are being tested in poplar includes ablation, gene suppression, and dominant negative mutants. 
Ablation of floral organs involves the regulated expression of a bacterial cytotoxin using a floral promoter 
from the species to be engineered. This approach targets expression of the cytotoxin to the tissues of the 
developing flower or floral buds, killing those tissues and thus preventing flowering. Engineering floral 
ablation in this manner is relatively easy to accomplish. However, very few if any genes are expressed 
absolutely exclusively in one cell type in plants. Promoters are more likely to be floral predominant 
rather than floral exclusive expression, thus raising the possibilit\ that the cytotoxin will also be 
expressed in non-floral tissues. Thus with the ablation approach, it will be necessar> to mitigate 
deleterious side effects in vegetative tissues prior to use. Another approach is to create dominant negative 
mutations in which a gene known to be essential for floral development is mutated in vitro to produce a 
protein that is stable in the cell but no longer active. Over-expressing that mutant gene produces an 
protein that interferes with the function of the wild type protein, causing sterility. 

One of the most active areas of research in plants is gene suppression bv RNA-interference (RNAi). 
RNAi is a natural system for regulation of gene expression in which small double-stranded RNA 
molecules interfere and thus suppress expression of endogenous genes. This can be engineered by 
expressing an invert::d repeat of a small fragment of DNA homologous to the target gene (De Buck et al. 
2001: Klahre ct al. 2002). It should be possible to create sterile trees by RNAi targeted at genes that are 
of essential lor floral development. Much is being learned about the genes regulating reproduction in 
trees, and this know ledge will offer opportunities for engineering sterility by manipulating the expression 
of nuclear genes as described above. Dr. Steven Strauss of The Tree Genomics, Biotechnology, and 
Breeding Program in the Department of Forest Science at Oregon State University, is conducting research 
on engineering sterility in GE poplar trees using floral genes in poplar as their model s\ stem. Table 1 
lists some of the genes that are known tc^ be iinoKcd in floral development in Populus species (as of 
May, 2004) 



154 



Table 1. Genes controlling flower development in Populus. 



Arabidopsis Gene 


Function in Arabidopsis 


Poplar Honiolog(s) 


AGAMOUS (AG) * 


Stamen & carpel identity 


PTAGl 
PTAG2 


APETALA3 (APS) * 


Petal & stamen identity 


PTD 


APETALAl (API) * 


Flower initiation; perianth 
identity 


PTAPl-l 
PTAPl-2 


LEAFY (LFY) 


Flower initiation 


PTLF 



*MADS-box gene, member of a large plant gene family that regulates the expression (transcription) of 
other genes. 

Field trials will be an important step in evaluating new GEO trees and in validating bioconfinement 
techniques such as engineered sterility. Strauss (2003) points out that a key issue in evaluating the safety 
of new GEOs and also of bioconfinement techniques should be the expected fitness consequences of 
escapes. That is, how great a risk to the environment or to people is posed by individual escapes, when 
bioconfinement techniques are not 100% effective. To be a risk, in most cases a transgene must amplify a 
great deal to have significant environmental impact. An individual GEO tree that has escaped into a 
natural population of millions of trees will have little of no impact on the overall makeup of that 
population, unless the transgene increases in frequency quickly. Small releases of GEOs thus have a 
major built-in safety buffer of scale in field trials. Unless the transgene provides a great differential 
benefit under strong selection pressure, the key force governing spread of the transgene will be genetic 
drift. Under genetic drift, a selectively neutral transgene is as likely to be lost from a natural population 
as maintained, and if not eliminated would require many generations to reach a frequency high enough to 
be a stable component of a natural population. Furthermore the maximum frequency that a neutral 
transgene might reach in a natural population will be lower, the fewer escapes from sterility that occur in 
the GEO population. This indicates that there should be a relatively high tolerance for incomplete 
sterility in field trials of GEO trees. If so, it may not be necessary to always accomplish 100% sterility to 
have an effective means of bioconfinement for transgenes. 



APPLICATION OF GENETIC ENGINEERING TO AMERICAN CHESTNUT 

The application of GE to American chestnut would involve transgenes for resistance to the blight 
resistance. A major issue with the release of GEOs with transgenes for resistance to disease and other 
pests is whether or not that will lead to resistance in the targeted organisms, as has occurred with the use 
of chemical pesticides. This is a particular concern with GEO trees which need to be able to rely on their 
transgenes for resistance over the course of many generations of the pest population. Several factors 
infiuence the development and containment of resistance in targeted pests. These factors include I) the 
genetic basis of resistance, 2) the initial frequency of resistance alleles in the target pest population, 3) the 
competitiveness of resistant individuals in the pest populations, and 4) the resistance management strategy 
employed. 

To ensure the long term usefulness of transgenes in GEOs for pest and disease resistance, it is necessary 
to have an effective plan to manage against the development of resistance to the transgene product in the 
pest populations. The key factors for a successful Resistance Management program include I ) knowledge 
of the biology and ecology of the pest, 2) low initial frequency of resistance genes in the pest populations, 
3) low survival of pests when they are heterozygotic for the resistance genes, limiting transmission and 
build up of resistance genes in the pest population, and 4) the establishment of nearby refugia where the 
GEO is not present to ensure that susceptible alleles remain in high frequency in the pest populations. 



.55 



Many different designs for refugia have been tested, but it most commonly it is recommended that 20% of 
the total area under cultivation, or crops, be planted as a refuge with non-GEO plants. Ihe exact position 
of the refugia \ary w ith the pest and crop species in question, but refugia ma> be embedded as rows 

within in the GEO field, or as borders or Blocks in the GEO 
field or even as a separate neighboring field. An alternate 
design includes a separate 20% Refuge that is sprayed w ith 
pesticide not related to the transgene product, to control the 
numbers of pests entering the GEO plots. 

It is important to note, that there should be much less concern 
about the break down of resistance in chestnut trees than 
would normalK be the case with crop plants. The sprouts that 
continue to arise from the roots of old wild chestnut trees will 
provide a large, continuously distributed natural refugia for 
the production of wild type blight spores (figure I ). The 
widespread presence wild-type sprouts that occurs throughout 
the natural range of American chestnut will provide the ideal 
refugia for preventing mutations in the Cryphoncctria fungal 
populations from getting the upper hand. Thus, blight 
resistance in chestnut, whether from the back-cross breeding 
program or from genetic engineering, should be relatively stable over time in the forest, and perhaps 
require less active management than disease resistance in annual crops rather than more. 



4, 


.jmm^i^yf, 




K'j 






\ 






Figure 1 . American chestnut tree 
with cankered trunk and new sprouts 
from the base that are not yet 
showing infection. 



RECOMMENDATIONS OF THE COMMITTEE 

After an exhaustive search which included interviews with experts from academia and the biotechnology 
industry, the committee reported the following major findings of its study: 

• Most GEOs will not require confinement. 

• The need for bioconfinement should be evaluated on a case-by-case basis. 

• The use of redundant biocontmement methods will be necessary in some cases. 

• Biological confinement of GEOs should be undertaken in the context of an integrated 
confinement system. 

• 1 he need for confinement should be considered at the beginning of the design of a GEO. and be 
part of the entire development process, not just at the end. 

The Committee on Biological Confinement of Genetically F>ngineered Organisms concluded its study 
with the following list of recommendations for the USDA. 

• The need for bioconfinement should be determined on a case-by-case basis. 

• The need for bioconfinement should be considered early in development of a GEO. 

• The level of confinement needed should be defined early in development of GEO. 

• The stringency of the integrated confinement system should reflect the predicted risk and severity of 

consequences of GEO escape. 

• Bioconfinement techniques should be relevant to the temporal and spatial scales of field release. 

• Confinement techniques should be tested experimentally. 



156 



• The phenotypes of novel GEOs should be compared with the progenitor organisms. 

• Due to the long times required, field tests of bioconfinement methods with trees should be started as 

soon as possible, even if tests must first be conducted with GE trees not requiring confinement, to 
produce data the needed for later releases. 

X., • Redundancy of methods can be used to improve confinement in high risk cases. 

• An Integrated Confinement System should be used, involving technical, organizational, and 

regulatory elements. 

• Methods should be developed to facilitate environmental monitoring for escapes, including easily 

identifiable markers, and sampling strategies. 

• Transparency and public participation should be important components in developing and 

implementing appropriate bioconfinement approaches. 

• The possibility of human error should be taken into account as a factor when determining 

bioconfinement methods and evaluating their efficacy. 

• The international effects of failures of confinement should always be considered. 

• International cooperation on the confinement of GEOs should be pursued. 

• More research is required on methods for biological confinement of GEOs. 

The committee went into some detail on the reasons why more scientific research is required. Research is 
needed to characterize ecological risks and consequences and develop methods and protocols for 
assessing the environment effects of confinement failure. Research is needed to develop reliable, safe, 
and environmentally sound bioconfinement methods. Research is needed to identify and develop 
methods and protocols to assess the efficacy of bioconfinement. Research is needed to identify economic, 
legal, ethical, and social factors that might influence the application of techniques, and their regulation. 
Research is needed to develop a better understanding of the dispersal biology of organisms targeted for 
genetic engineering and release. Research is needed to develop a better understanding of how species 
become invasive. Finally more research on risk assessment and safety management specific to GE trees 
(as a follow up to Lu et al 1999) are needed. 

In addition, the Committee prepared a template for risk evaluation to be used in the process of planning 
GEOs, including in decision making on the need for and methods of biological confinement. This risk 
evaluation template could be a good tool for the National ParK Service to use in evaluating the use of 
specific GE trees in ecosystem restoration and enhancement projects. 



LITERATURE CITED 

Connors, B. J, Laun, N. P, Maynard, C. A, Powell W. A. 2002a. Molecular characterization of a gene 
encoding a cystatin expressed in the stems of American chestnut {Castanea dentata). Planta 215 (3): 5 1 0- 
514. 

Connors, B.J., M. Miller. C.A. Maynard and W.A. Powell. 2002b. Cloning and characterization of 
promoters from American chestnut capable of directing reporter gene expression in transgenic 
Arabidopsis plants. Plant Science I63(4):771-78l. 



157 



De Buck, S., Van Montagu, M., Depicker, A. 2001. "Transgene silencing of invertedly repeated 
transgenes is released upon deletion of one of the transgenes involved." Plant Molecular Biology 
Reporter, 46: 433-445. - \ . 

Kirk, T.K., Carlson, J.E. et al., 2004. Biological Confinement of Genetically Engineered Organisms, The 
National Academies Press, Washington, DC, 255 pages. 

Klahre, U., Crete, P., Leuenberger, S.A., Iglesias, V.A., Meins Jr.. F. 2002. "High molecular weight 
RNAs and small interfering RNAs induce systemic posttranscriptional gene silencing in plants." 
Proceedings of the National Academy of Sciences. 99: 1 1981-1 1986. 

Lu M.Z., Han, Y.F.. Du S.M. 1999. "Risk assessment and safety management of genetically engineered 
trees." Forest Research, 12: 325-33 1 . 

McLean, M.A., Charest, P.J. 2000. "The regulation of transgenic trees in North America.." Silvae 
Genetica, 49: 233-239. 

Merkle, S.A., Dean, J.F. 2000. "Forest tree biotechnology." Curr Opinion in Biotechnology. 1 1 : 298- 
302. 

Pena, L., Seguin, A. 2001 . Recent advances in the genetic transformation of trees. Trends in 
Biotechnology, 19: 500-506. 

Sederoff, R. 1999. "Building better trees with antisense." Nature Biotechnolog> 17:750-751, 

Slavov, G. T., DiFazio, S. P. Strauss, S. H. 2002. Gene flow in forest trees: From empirical estimates to 
transgenic risk assessment. In: Gene Flow Workshop, The Ohio State University, March 5 and 6, 2002, 
pp. 113-133. 

Strauss, S.H. 2003. Regulating Biotechnology as though Gene Function Mattered. BioScience. 53: 453- 
454. 

Strauss, S. H., Rottmann, W. H., Brunner, A. M. & Sheppard, L. A. 1995. Genetic engineering of 
reproductive sterility in forest trees. Molecular Breeding 1 : 5-26. Tang and Tian, 2003 

Trifonova, A., Atanassov, A. 1996. "Genetic transfomiation of fruit and nut species." Biotechnology and 
Biotcchnological I qiiipment, 10: 3-10. 

www.isb.vt.edu/cfdocs/fieldtestsl.cfm. "Field test releases in the USA" (Information Systems for 
Biotechnology, Virginia Tech, Blacksburg. VA 2003). Accessed, May 22, 2003. 



158 



Steiner, K. C. and Carlson, J. E. eds. 2006. Restoration of American Chestnut To Forest Lands - Proceedings of a 
Conference and Workshop. May 4-6, 2004, The North Carohna Arboretum. Natural Resources Report 
NPS/NCR/CUE/NRR - 2006/001. National Park Service. Washington, DC. 

NATIONAL PARK SERVICE POLICIES: HIGHLIGHTS FROM A WORKSHOP ON 
GENETICALLY MODIFIED ORGANISMS (GMOs) IN PARK LANDS 

^'~ William A. Leilis 

National Park Service, Chesapeake Watershed CESU, 301 Braddock Road, Frostburg, MD 21532, USA 

(wlellis@al.umces.edu) 



Abstract: A workshop was sponsored by the National Park Service (NPS) to begin developing policy 
regarding use of genetically modified organisms (GMOs) within NPS resource management programs. 
GMOs were defined as organisms that contain gene combinations that do not occur naturally and were not 
created through traditional breeding practices. There are no current NPS policies that specifically prohibit 
use of GMOs on park lands. Therefore, for the purpose of the Chestnut Workshop, all technologies will 
be considered as potential tools to help restore American chestnuts to lands where they once occurred. 

Key words: GMO / Workshop / NPS / Policy 



INTRODUCTION 

In April 2004, the National Park Service (NPS) sponsored a workshop to begin the process of developing 
policy regarding use of genetically modified organisms (GMOs) within NPS managed lands. The 
objectives of the workshop were to educate a core group of NPS personnel on various GMO issues, 
discuss where GMOs are currently being used on NPS lands, and identify unclear or controversial issues 
that would need to be resolved before policy development. 

The GMO Workshop was structured much the same as the Chestnut Workshop, with two days of invited 
talks followed by a day of deliberations by NPS staff to begin drafting the major elements of a GMO 
policy. Invited talks covered basic concepts of GMO technology, current use of GMO products in North 
America, GMO products in development, current and potential use of GMOs in National Parks, potential 
environmental and social risks associated with GMOs. and a review of NPS policies that relate to GMO 
use in resource management programs. The Chestnut Workshop was scheduled to occur after the GMO 
Workshop in case policy described or developed at the GMO Workshop would clearly preclude use of 
genetically engineered products in an NPS chestnut restoration program. 



WORKSHOP OUTCOMES 

NPS policy is primarily contained within the publication Management Policies last revised in 2001 . 
Policy can be supplemented or amended between revisions through formal issuance of a Directors Order. 
Management Policies 2001 contains only brief direct references to GMOs, but does contain substantial 
guidelines on use of biological products to attain resource management goals. Those include: 1 ) guidance 
to restore extirpated native species using the closest available genetic material (4.4.2.2); 2) ability to 
introduce an e.xotic species in rare situations to meet specific management objectives (4.4.4.1); 3) ability 
to use a hybrid, subspecies, or improved variety where the natural variety cannot survive human-altered 
environmental conditions (4.4.4.1 ), and: 4) the ability to use bioengineered products for exotic pest 
management (4.4.5.4). There are no specific prohibitions against using GMOs within National Parks. 



59 



Genetically engineered agricultural crops, primarily herbicide tolerant and insect resistant com and 
soybeans, are currently used in several agricultural lease programs by militar> and historical parks that are 
mandated to preserve farmed fields within the historic landscape. These parks generally lease fields to 
local farmers through cooperative use programs. GMO crops are often preferred because thev are 
believed to decrease total use of pesticides, enable use of less toxic herbicides, and reduce fuel and labor 
costs. In addition, the use of GMO soybeans is so prevalent in U.S. agriculture that it is often difficult for 
farmers to obtain non-GMO seed in the commercial market. , 

A second area where GMO products are currently used in NPS management programs is recombinant 
wildlife vaccines such as rabies vaccines for coyotes and raccoon, canine distemper v accines for black- 
footed ferrets and fo.\, and equine West Nile Virus vaccine for horses and mules. It is possible that other 
GMOs are being introduced on NPS lands without parks being aware they contain engineered genes 

A proposed working definition of GMOs was those organisms that contain gene combinations or gene 
sequences that do not occur naturally and were not created through traditional breeding practices. This 
generally refers to technologies that remove a small number of genes from one or more donor organisms 
that are inserted into a receiving organism, often across taxonomic groups. For NPS policy purposes, this 
would not include hybrids produced through artificial breeding or products of GMOs that do not contain 
viable genetic material, such as killed vaccines. There was also discussion on the benefits and limitations 
of categorizing all GMOs as exotic species for policy purposes. 

The general feeling among workshop participants was that existing NPS policy would allovs introduction 
of a GMO into a park if it met clearly defined management objectives and all feasible and prudent 
measures were taken to minimize risk or harni to other natural and cultural resources. Other policy 
considerations discussed included: I ) a prohibition against using GMOs for purelv aesthetic purposes: 2) 
a prohibition against using GMOs if doing so would jeopardize park objectives or pose risk to human 
health or safety: 3) a prohibition against growing GMO crops that produce phannaceuticals: 4) stringent 
risk & benefit analysis before considering introduction of a GMO. including full NEPA compliance: 5) 
consideration of possible gene fiow to areas outside the park, particularly if certified organic farms are 
nearby: 6) monitoring use of GMOs outside of park boundaries for possible impact to park resources: 7) 
approval of GMO use on a case-by-case basis, and: 8) annual reporting of all GMOs used or released on 
NPS managed lands. 

CONCLUSIONS 

The process of constructing NPS policy on use of GMOs in park resource management programs has just 
begun, and will be developed and formalized over the coming years. At present, there are no specific 
prohibitions against using a GMO in park programs. The general feeling of workshop participants was 
that a GMO might be deemed acceptable if it is the best available product to meet a specific management 
goal, is generally considered safe by the scientific community, is accepted by the public, and is approved 
through the NEPA process. Therefore, for the purposes of the Chestnut Workshop. an\ lechnologv can 
be discussed for possible incorporation into an NPS restoration program, including gencticalK modified 
trees and fungal pathogens. 



160 



Steiner, K. C. and Carlson, J. E, eds. 2006. Restoration of American Chestnut To Forest Lands - Proceedings of a 
Conference and Workshop. May 4-6, 2004, The North Carolina Arboretum. Natural Resources Report 
NPS/NCR/CUE/NRR - 2006/001, National Park Service. Washington, DC. 

PLANTING TRIALS OF AMERICAN CHESTNUT IN CENTRAL APPALACHIAN 

MOUNTAINS 

3 _^ Timothy R. Phelps'. Kim C. Steiner', Chien-Chih Chen", and James J. Zaczek^ 

School of Forest Resources, The Pennsylvania State University, 
University Park, PA 16802 (phelpstfa)psu.edu) 
" Department of Horticulture, The Pennsylvania State University, University Park, PA 16802 
^ Department of Forestry, Southern Illinois University, Carbondale, IL 62901-441 1 



Abstract: Field planting methods for American chestnut were examined in three separate trials to develop 
guidelines for the anticipated establishment of blight-resistant hybrid American chestnut into Appalachian 
forests. A direct-seed tree shelter test examined the effect on height growth of using five-foot tall tree 
shelters, vented or unvented, and no tree shelter treatments. A containerized/nursery stock test examined 
the effect on height growth of two nursery stocks (1-1 and 1-0) and greenhouse raised containerized stock 
in two container sizes (40 and 10 cu. in.) where half of each of the containerized stocks were given a 2.5- 
foot tree shelter. The direct-seed tree shelter test and the containerized/nursery stock test were located 
adjacent to each other at two sites. In addition, a site evaluation test examined the suitability of planting 
American chestnut at seven forest sites that formerly supported chestnut. In the absence of deer 
browsing, 5-foot tree shelters, vented or unvented, and 2.5-foot tree shelters, had no significant advantage 
over unsheltered treatments for seedling height after three field seasons. However, tree shelters were 
absolutely critical where deer browsing was frequent. 1-0 nursery stock did not grow significantly better 
than older 1-1 stock beyond the first field season, indicating that the extra year in the nursery was not 
necessary. Container size had no significant effect on growth rate. Planted seedlings competed well 
when natural regeneration was reset to ground level mechanically. Height and survival were 
unacceptably low for successful regeneration in all site evaluation trials, probably because of our 
inexperience in direct-seeding this species and intentionally casual approach to controlling competition 
and access by deer. 

Keywords: direct-seed / container stock / nursery stock / tree shelter / regeneration / browse / transplant / 
seedlings / reforestation / restoration 



INTRODUCTION 

Field planting methods for American chestnut were examined to develop guidelines for the anticipated 
establishment of blight-resistant hybrid American chestnut into Appalachian forests. Much of what is 
known about American chestnut silviculture and regeneration ecology is derived from observations and 
studies that were carried out before the blight and the advent of forestry research as we now know it. 
Prior to the blight (Cryphonectria parasitica), American chestnut was found on gentle to steep slopes in 
mixtures with pines, oaks, and other hardwoods. The species avoided limestone-derived soils or bottoms 
with wet, cold, or shallow soils (Buckhout 1 896. Zon 1904). Much of the reproduction was of coppice 
origin as most nuts were consumed by wild animals, livestock, and man (Buckhout 1896, Zon 1904). 
Seedlings that were able to germinate grew rapidly and formed long, vigorous, and wide spreading root 
systems similar to oak (Toumey and Korstian 193 1 ). Regeneration was likely favored by widespread 
clearcutting and wildfires as chestnut was not as shade tolerant compared with beech, maple, and other 
potential competitors (Toumey and Korstian 193 1 ). 

Based on the authors' experience with oak, a relative of chestnut, trials were established to examine 
planting methods previously used for oak with the assumption that the results would be similar. Three 



161 



trials were performed: 1 ) a direct-seed tree shelter test, 2) a containerized/nursery stock test, and 3) a site 
evaluation test. 



STUDY ONE 



Protection from white-tailed deer {Odocoileus virginianus) is essential to establishing healthy forest 
seedlings in many parts of central Appalachia. When valuable disease-resistant American chestnut 
genotypes become available as planting stock for restoration, extraordinary protection measures such as 
use of fencing or tree shelters will be warranted. Tree shelters are relatively effective in protecting 
seedlings from deer until the tree grows above the deer browse line. The shelters also purportedK provide 
favorable growing conditions by moderating environmental extremes. However, shelters reduce light 
intensity, physically limit display of leaves to sunlight, and can increase temperatures around the seedling. 
Tree shelters often induce seedlings to grow faster in height than in diameter, and the resulting trees are 
typically spindly and susceptible of falling over if not supported. Shelters can ha\e warmer internal 
temperatures compared to ambient conditions, which can accelerate bud break in the spring or delay 
hardening-off in the fall rendering trees susceptible to frost damage or winter injur>'. To prevent this. 
Tree Pro (West Lafayette, Indiana), a manufacturer of tree shelters, produces a vented shelter designed to 
maintain cooler temperatures and thus allow the seedling to acclimate more normally in autumn. 

In 1997, a planting of direct-seeded American chestnut was established at Stone Valley (SV), The 
Pennsylvania State University's Experimental Forest in Huntingdon County, Pennsylvania. The test was 
designed to measure the effect on height growth of vented and unvented. five-foot-tall tree shelters (Tree 
Pro) vs. no tree shelter. For tree shelter treatments, American chestnut seed (obtained from Philip Lunde. 
Galesville. Wisconsin) was planted 1 inch below the soil surface and a vented or unvented tree shelter 
was erected over the planting spot. Shelters were pressed into the soil I in. Seeds for the unsheltered 
treatment were planted only within a seed protector, a 6 in. long piece of pre-split I in. diameter PVC pipe 
inserted into the ground to inhibit seed predation by small mammals. Fifty American chestnut seeds were 
planted for each treatment in a randomized complete block design. The study was established in a recent 
sheltcrwood harvest area (50 percent basal area removed) and a six-strand electric fence was erected 
around the entire site to provide protection from deer. Fortunately, this area has a ver\ low deer 
population, and the need for protection from deer was minimal. Nonetheless, fencing allowed for a 
comparison of the shelters' effects on chestnut seedling growth without potential confounding from the 
effects of deer browsing. Native chestnut sprouts occur immediately next to the study area, so the site 
may be considered suitable for this species. 

As anticipated, seedlings were significantly taller (P<0.05) in shelters compared to those left unsheltered 
for the first two years, but there was no statistical difference in height between the t>pes of shelter in an) 
year. Trees with no shelter began to catch up during the third growing season, to the point that there was 
no significant difference between treatments, but their mean height was still lower by about one foot 
compared to the sheltered treatments. Over the next three years, unsheltered trees became substantially 
taller than sheltered seedlings, although not significantl\ so. h\ nearK one. two. and three feet for each 
respective year. Certainly had there been deer pressure at this site unsheltered trees would ha\e had much 
more difficulty in getting established and growing be\ond the deer-browse line. After se\en \ears of 
growth, the average heights are 21.1, 17.7, and 15.1 feet for unsheltered, ventcd-shelter, and unvented- 
shelter trees, respectively, and the tallest tree is 29.2 feet (unsheltered). This confimis the excellent 
suitability of this site for chestnut growth, not to mention the rather startling potential of this species to 
grow v\ell in forest plantations. Blight is beginning to appear in the plantation. 

A replicate test was begun in 1998 at the SV site and at a site in Tuscarora State Forest (TSF). about 70 
miles south in Periy County, Pennsylvania. However, bears continually ravaged the SV planting by 



162 



tearing apart the shelters, and data collection was discontinued in 2002 because the plantation was too 
disrupted to provide meaningful results. Fortunately, the bears concentrated their efforts in this replicate 
and did minimal damage to the adjacent 1997 trial. The TSF site differed primarily in the amount of 
sunlight reaching the trees, as this plantation was established in a clearcut (rather than a shelterwood). 
This site also had an electrified fence, but the fence was largely ineffective and there was very heavy deer 
browsing on vegetation within the fence. 

As in the 1997 SV test, initial heights at TSF were greater with tree shelters than without, and there was 
no significant difference in mean height between shelter types until after the fifth growing season when 
the trees had grown well above the tops of the shelters and trees in vented shelters were taller. In sharp 
contrast to the SV test, trees without shelters at TSF have never grown past competing vegetation or 
above the height of deer because of continual deer browsing. After six years of growth, the average 
heights are 2.2, 14.8, and 12.9 feet for unsheltered, vented-shelter, and unvented-shelter trees, 
respectively. The sheltered treatments are on par with parallel treatment means at SV of the same age. 

The results of this study show that shelters do not benefit, and may even retard, the growth of American 
chestnut in forest plantations in the absence of deer pressure . Where deer were abundant, five- foot tree 
shelters were far more advantageous than no protection. No significant difference in growth was detected 
between trees in vented as opposed to unvented tree shelters. 



STUDY TWO 

While direct seeding will likely be the most efficient and cost-effective planting method for reestablishing 
chestnut into the central Appalachian forests, planting seedlings ensures higher survival rates per seed and 
permits greater control over final tree placement. Transplanted seedlings may also be competitively 
superior against the understory vegetation often encountered on forest sites. 

Evaluations of the performance of containerized and nursery stock were performed at the SV and TSF 
sites adjacent to the previously mentioned direct-seeded, tree-shelter studies. The purpose of this study 
was to compare the growth of seedlings grown in two sizes of containers and raised in a greenhouse for 
three months (mid-February thru mid-May) and seedlings that had been grown under standard forest 
nursery conditions at a nearby Bureau of Forestry nursery for one or two years. The large and small 
containers measured 10 inches in length by 2.5 inches in diameter (40 cu. in.) and 8.25 inches in length by 
1.5 inches in diameter (10 cu. in.), respectively. One hundreci of each type of containerized stock was 
planted at each site, half of which also had a 2.5-foot shelter for limited deer protection. The nursery 
stock consisted of 20 each of 1-0 and 1-1 (one year in seedbed, one year in transplant bed) stock at each 
site. All seeds were obtained from the same source as the previous study. Each site was established in a 
completely randomized design in early- to mid-May of 1999. Existing vegetation was cleared to the 
ground with a brush cutter in an effort to reset competition. An auger with a 4-inch bit was used to plant 
the seedlings. 

All planting stock transplanted well at both sites with better than 90 percent survival across all treatments 
and sites after two years in the field. More individuals began to die in the third year as other limiting 
factors (e.g., deer pressure and competing vegetation) became increasingly severe. 

At SV, heights of the 1-1 and 1-0 nursery-grown seedlings were significantly greater than those of 
younger, containerized seedlings after the first year in the field. There was no significant difference in 
height between 1-1 and 1-0 seedlings after two years in the field, which means that the extra year of 
growth in the nursery was not of practical advantage. The extra resources of the nursery-grown seedlings 
(initially larger root and shoot systems) were still providing a height advantage to those seedlings 



163 



compared to the younger, containerized material through five field seasons. The only exception was the 
large-container / no-shelter treatment, which was not significantly shorter than seedlings grown from 1-0 
nursery stock after the fifth year; mean heights after five growing seasons were 9.4, 8.2. and 6.5 feet for 
I - 1 , 1-0, and large-container / no-shelter treatments, respectively. 

Containerized stock at SV was fairly consistent in height across treatments through three field seasons, 
but by the end of the fourth season, seedlings grown without shelters began to outgrow their sheltered 
counterparts. The height advantage of unsheltered seedlings was statistically significant by age five. 
These results precisely mirror those of Study 1 with 5-foot tree shelters. After five \ears of growth, there 
was no significant height difference between seedlings started in different-sized containers and planted 
without shelters, but percentage survival was slightly higher with the seedlings that were started in the 
larger container (86 vs. 78 percent). In general, container-grown seedlings at SV have lagged in growth 
about a year behind the one-year-older 1-0 nursery stock, but not quite two years behind the two-years- 
older 1-1 stock. 

At TSF the story was again vastly different due to significant deer pressure. While the nursery stock held 
height advantages over containerized material after the first growing season. 1-1 and 1 -0 stock exhibited 
only meager growth in following years. Because all l-l and 1-0 seedlings were unprotected b\ shelters, 
the deer were able to continually browse new growth, and those seedling treatments ha\e essentially 
failed at TSF. After five growing seasons, survival was only 40 and 45 percent for 1-0 and 1-1 nursery 
stock treatments, respectively. 

Seedlings started in containers and provided with tree shelters have done better at TSF. Sheltered stock 
started in large containers have been taller than seedlings in other treatments since sometime in the second 
growing season, although sheltered seedlings started in small containers began to catch up by age five. 
However, with few exceptions, seedlings have never grown much be\ond the height of the shelter itself 
(2.5 feet), and the mean height of all seedlings in shelters was only 3.3 feet after five field seasons, in 
addition to the deer pressure, these seedlings had to contend with an abundant cohort of yellow -poplar 
seedlings that became established simultaneously with the installation of this study. Yellow-poplar is a 
famously competitive species on good sites, but some of the chestnut seedlings that managed to escape 
the deer are competing fairly well with the yellow-poplar, typically Just behind them in height. We will 
continue to watch these with great interest. 

Overall, this study shows that transplanting of both nursery and greenhouse-grown, containerized stock 
can be accomplished with great success, but only if deer browsing is not a factor. In the absence of 
browsing, planted American chestnut seedlings can compete well with surrounding natural regeneration 
through five field seasons if that regeneration is reset mechanically at the time of field planting. If deer 
browsing is not a factor, seedlings grow better without tree shelters than with. There appears to be no 
advantage to using 1-1 nursery stock, which is costly to produce and difficult to plant compared to 1-0 
nursery stock. 



STUDY THREE 

In our latest phase of experimental chestnut plantings, we are examining the suitability for American 
chestnut of a range of native and relatively undisturbed forest sites by attempting direct-seeded 
establishment of chestnut plantations. The Pennsylvania Chapter of the American Chestnut Foundation 
provided seed for this study. Seven sites were established with 50 seeds each in 2001 and this was 
replicated in 2002. The sites var\ in soil t\pe. ele\atioii. aspect, competing ground vegetation, and 
undoubtedly other important respects. All plantations were established in fenced areas that had recently 



164 



been harvested, or in areas that were fenced to encourage regeneration where little vegetation existed due 
to over-browsing by deer. 

Two of the sites were located approximately 80 miles south of State College, PA, in Tuscarora State 
Forest. "Eby Ridge" has a Hazleton extremely stony, sandy loam soil that consists of a deep, well- 
drained, strongly acid soil that formed under sandstone residuum. It sits on a south-facing slope at 1 237 
ft. It had little competing ground vegetation after the first year of establishment, but the 2001 trial later 
developed considerable hardwood competition. "Dead End Road" consists of the same soil type, but sits 
on the east side of a ridge top at 2026 ft. and has considerable competition with a thick carpet of 
Vacciniuni. While this site is fenced, it still has a considerable presence of deer. 

"Deep Hollow" is located approximately 40 miles east of State College in Bald Eagle State Forest. The 
soil type is delineated as Dystrochrepts bouldery great group, but it appears very similar in texture and 
stoniness to Hazleton. In fact, this site bears a very strong resemblance to Eby Ridge in most regards. 
Deep Hollow sits on a south-facing slope at 1302 ft. and had only a small amount of competing 
Vaccinium on the 2001 trial, but a greater density on the 2002 trial. 

Four sites are located within Rothrock State Forest south of State College. "Galbraith Gap" has a Laidig 
extremely stony loam soil that consists of a deep, well-drained, strongly acid soil that formed under 
sandstone and siltstone alluvium. It sits on a south-facing slope at 1930 ft. and has a thick carpet of 
Vaccinium. Here again, although this is a fenced area, there is a considerable presence of deer. "Pine 
Swamp Road" has a Hazleton-DeKalb soil type very similar to the Hazleton series. It sits on an east- 
facing slope at 1442 ft. and has a thick carpet of hay-scented fern. "Owl Gap" has a Buchanan extremely 
stony loam soil that consists of a very deep, moderately well-drained, slowly pemieable, very strongly 
acid soil. Buchanan series soils formed in colluvium on mountain footslopes, sideslopes, and in valleys 
that were weathered from acid sandstone, quartzite, siltstone. and shale. It sits on a north-facing slope at 
1403 ft. and has a thick carpet of hay-scented fern. "Spruce Mountain" also has a Buchanan soM type. It 
sits on a south-facing slope at 1593 ft. and has some competition with grasses, forbs, and regenerating 
sassafras and red maple. 

All sites were mechanically cleared before planting with a brush cutter to reset competing vegetation. 
Radicles had emerged from most seeds while in storage and were cut back to facilitate planting in seed 
protectors (described above) that were used for protection against small mammals. The competition 
described above was present at the time of data collection in late September 2001 and was similar through 
2003. 

2001 Trial 

After three years in the field, seedlings at Owl Gap were tallest (20.0 inches) and had the best survival (66 
percent) compared to the other sites. Spruce Mountain was second best in height ( 1 7.4 inches), but 
ranked last in survival (26 percent). Eby Ridge and Deep Hollow were close behind in height ( 1 5.4 and 
15.0 inches, respectively), but had lower survival (38 and 47 percent, respectively). The rapid 
reemergence and density of the hay-scented fern at Pine Swamp Road seemed to be too much for the 
chestnut seedlings to develop much height (7.5 inches) while survival slowly declined to 46%. Finally, 
the intensity of deer browsing at Dead End Road and Galbraith Gap made the growth and survival of any 
edible plant almost impossible, so work at these two sites was discontinued. However, growth was 
unacceptably poor at all sites, and even the best survival rate (66 percent) is borderline by our standards 
of a successful forest plantation. 



165 



2002 Trial 

Owl Gap, Eby Ridge, and Spruce Mountain had the tallest trees (20.0. 18.7, and 14.0 inches) at the end of 
the second year. Mean seedling height at Deep Hollow was substantially less (9.4 inches), and the 
seedlings under the dense hay-scented fern cover at Pine Swamp Road reached a mean height of only 5.5 
inches. Mean heights after two years in the field in this trial were similar to mean heights after three 
years in the field in the 2001 trial, but no site is exhibiting the growth that chestnut is capable of. Survival 
at all sites was unacceptably poor, with Owl Gap having the highest at only 48 percent. 

These trials are still in their infancy, and our methods need to be refined until we can get better survival 
and more acceptable early height growth. Other than deer browsing, competing vegetation appeared to be 
one of the greatest factors affecting height at all sites, and resetting it mechanically back to the ground 
level apparently did not give direct-seeded chestnut sufficient early advantage for successful 
establishment. These plantations would have been failures if the goal had been to restore American 
chestnut to these sites. Plans to continue these studies with additional trials utilizing chemical control of 
competition and planted seedlings are underway. 



CONCLUSIONS 

After examining several planting methods we still believe direct seeding will be the most efficient and 
economical method of reestablishing American chestnut when large numbers of seed become available, if 
competition and deer browsing can be adequately controlled or avoided, as these appear to be absolutely 
critical factors when planting chestnut from seed. The growth and survival of seedlings at SV in Study 1 
demonstrate the remarkably vigorous potential of American chestnut in forest plantations, but the failures 
at TSF and in Study 3 show that this potential is fragile if conditions are not right. With a small amount 
of seed or available land area, one may well chose to use either bareroot or containerized nursers stock to 
ensure greater survival. We found 1-0 nurser>' stock to be just as good as 1-1. thus there is no real need 
for the extra year at the nursery. Containerized material had an excellent success rate but was costly to 
produce. 



LITERATURE CITED 

Buckhout, W.A. 1 896. Chestnut Culture for Fruit. Tlie Pennsylvania State College Agr. Exp. Sta. Bull. 

no. 36. 

Toumey, J.W., and C.F. Korstian. 193 1 . Seeding and planting in the practice of forestry . John Wile\ & 
Sons, Inc., New York. 507 p. 

Zon, R. 1904. Chestiiut in southern Maryland. USDA Bureau of Forestry Bull. No. 53, 31 p. 



166 



Steiner, K. C. and Carlson. J. E, eds. 2006. Restoration of American Chestnut To Forest Lands - Proceedings of a 
Conference and Workshop. May 4-6, 2004, The North CaroHna Arboretum. Natural Resources Report 
NPS/NCR/CUE/NRR - 2006/001, National Park Service. Washington, DC. 

PLANTING TRIALS WITH AMERICAN CHESTNUT 

IN SOUTHERN APPALACHIAN FORESTS (transcript of presentation) 

David Loftis 

Southern Research Station, USDA Forest Service, Bent Creek Experimental Forest, 

1577 Brevard Road, Asheville, NC 28806 USA (dloftis@fs.fed.us) 

INTRODUCTION 

We are really just getting started in our studies with American chestnut in the Forest Service. I don't have 
as many results to show as the folks at Penn State. I do first want to give some credit to the people who 
have really done most of the work in this study - particularly Chuck Rhodes, who was at the University 
of Kentucky at the time we started this and is now with the Forest Service out in Colorado and Jeff Lewis, 
who is a silviculturist in the Morgan Ranger District of the Daniel Boone National Forest. 

I am going to start off with this profound statement, 'planted chestnut is only going to make it if it beats 
its competition.'' The name of the game in forest regeneration is beating the competition. There are a 
couple ways to beat the competition. One way, from a silvicultural standpoint, at least, is that you can 
control the competition. We have all been exposed to weed eaters and herbicides and probably other 
techniques we can use. I wonder if in some cases, at least on a broad scale, whether an approach that 
relies strictly on control of competition is logistically and economically feasible. It may be, it may not be, 
I don't know the answer to that. This may just be a question of how much value we put into restoring 
chestnut. There may be other ways, other approaches, that we can take that might allow us to use less 
competition control and perhaps avoid some of the difficulties and expense with trying to reintroduce 
American chestnut on say a 100,000-acre piece of ground. 

WHAT DO WE KNOW ABOUT CHESTNUT? 

Well, we know something about its distribution. I was asked to talk specifically about the southern 
Appalachians. We know about where chestnut grew, and we also know that a variety of species now 
occupy the space once occupied by chestnut. But its function has probably not been entirely replaced in 
these ecosystems in the southern Appalachians, or elsewhere. Why haven't those functions been 
replaced? It was the most numerous, most abundant tree in the southern Appalachians. As an individual 
species, it was the most numerous. It was an important food source for many animals. One thing learned 
recently is that it was extremely important as long-term durable coarse woody debris in both terrestrial 
and aquatic systems. That's something that's disappearing now, and I think some people are concerned 
that we don't have a good replacement. And of course it was commercially valuable and had great utility. 
So these are some of the reasons why we want to regenerate chestnut and restore it. 

We also know that chestnut was a large tree. We know that it grew over a broad range of sites and soils. 
It certainly grew on mesic sites, where it reached its greatest development but it also grew and was 
actually more dominant on sub-xeric to even xeric sites. It grew over a broad elevational range, as well, 
in the southern Appalachians. We know that its wood was durable and extensively used. It regenerates 
well from stool shoots and from seeds, which are borne regularly in abundance. The rate of growth is 
very rapid, being greater than any other hardwood in the region. 



167 



i> 




Unfortunately, as has been already pointed out at this 
meeting, by the time we began to get systematic 
scientific investigations going in the eastern United 
States, the chestnut was on its way out. Figure 1, as an 
example, is a picture taken in 1905. This photo was 
taken on Bent Creek, and this guy was stud\ ing what 
was replacing chestnut. This picture was taken probabK 
just a few hundred yards from the North Carolina 
Arboretum. We really never did learn much about the 
silvical or ecological characteristics of American 
chestnut before its decline. 

Figure 1 . Early study on American chestnut 
regeneration. Bent Creek, 1905. 



WHAT DO WE NEED TO KNOW? 

The work of Ayers and Ashe (1905) restates what we discussed earlier -chestnut grew very rapidly. 

What they observed as they traversed the woods is that yellow-poplar also grew verv rapidiv . And I 

particularlv like their 
description of yellow-poplar 
and chestnut "freely seeding' 
into openings created after 
cutting (Figure 2). The young 
poplars often vastly 
outnumbered the chestnuts, 
but the chestnut grew faster 
and overtopped the poplar and 
suppressed it. which is an 
interestinsz observation. 



DR.\IN.\GE BASINS. 51 

The moro rapid height g^rowth of the cheatmit in hir<,'c ineasuro ix-counts for 
the .-Clint roprodiiction of tht? yellow |K>pl«r in culled woods, for jjopliir imd 
clie.^tniit lioth freely seed such <>ptnin<.'<. the youiij,' poplarr, often outnumlfering 












^^^ 


....- 


_. 
















-^ 


^■"^ 








ao 

60 

*o 






/ 
















/ ••'^' 














J/ 






























Fio. 2.— Ctllv 

the che.stiiuts, but 
suDoresses it. 


!0 ■to (lO ao lOO 120 u 

eh»hoTrinK thermtcaiif hciehi int>wth of yell*^w poplar and 

the chestnut grows far mure rapidly, < 


'Oyeara 

chiMnni or (tomJ aoll. 

avertopa the poplar, and 



But Ayers and Ashe were not 
the only people around during 
the period before chestnut 
began to decline. Another 
\erv keen observer was E. H. 
Froth ingham. He came 
through the Asheville area in 
the early part of the century, 
actually about 1 9 1 4- 1 9 1 5, as 
the national forests were 
being set up pursuant to the Weeks Act. Frothingham also later became the first director of what is now 
the Southern Research Station. He was also Director of the Appalachian Forest Ivxperiment Station for 
the U.S. Forest Service from 1921 to 1934. He did a report (Frothingham 1917) on the cut-over areas of 
the southern Appalachians, from North Carolina to West Virginia. It was a very complete report. It was a 
thorough observational study, looking at many different cuts and what was becoming reestablished in 
those cuts. 



Figure 2. Yellow-poplar and American chestnut competition (Ayers 
and Ashe 1905). 



168 



But while Ayers and Ashe seemed to suggest that new seedlings were a viable source of regeneration in 
forest openings, that was not what Frothingham routinely observed. In fact, in only one case in all of his 
observations did he indicate that he thought that the source of successful regeneration was coming from 
seedlings. It was almost always from sprouts. There are some contemporary scientists (e.g., Billo, 
Paillet, McNab) who have painted a similar picture of the natural regeneration ecology of chestnut, 
indicating that it might not freely regenerate from new seedlings, at least over all sites. 

Frothingham (1917) also stated that: "'Observational studies of cut-over areas have the great disadvantage 
of missing the most important stages in the reproduction on cut-over areas, and especially the conditions 
prevailing immediately before and after the year of cutting." 

This was a criticism he made of his own study and other observational studies. In observational studies 
you come in after the fact and maybe look at a time series of different cuts on different sites. This 
approach misses some of the most important information about factors that ultimately influence species 
composition. And over the last 40 years we have come to realize that one of the most important things 
that we need to know about species is their characteristic regeneration strategy - what is the source of 
successful reproduction? Reproduction comes from a finite number of sources: from new seedlings 
established after disturbance, from advance reproduction that persists through disturbance, and from 
sprouts from stumps or roots that also persist through disturbance. 



WFLAT DO WE KNOW ABOUT CHESTNUT REGENERATION? 

We really don't know for sure where successful natural regeneration of American chestnut came from. 
Although we know that sprouts were successful, we don't necessarily know that seedlings were often 
successful. Can chestnut become established after disturbance and grow rapidly enough to compete 
successfully or does it have to persist through disturbance as advance reproduction, like the oaks and 
hickories and most other species that we have in the southern Appalachians? Perhaps it is like the oaks, 
which don't grow rapidly from small advanced reproduction as black cherry does, just as an example. 
Like the oaks, it might require a larger root system to sustain rapid height growth and to compete 
successfully. If it does require a large root system, what kind of light regime is needed to create or 
develop that large root system, and how long does it take for it to develop under those conditions? 

I think the last question is extremely important. Does it behave consistently; is its regeneration strategy 
the same everywhere? Is it the same on xeric sites as it is on mesic sites? Is regeneration the same in 
Massachusetts as it is in the southern Appalachians? I don't think we really know the answers to these 
questions. It could well be that we'll find multiple strategies necessary in order to regenerate chestnut. 



PRELIMINARY FINDINGS 

In our initial study, which is only a couple of years old now, we decided to look at regeneration as well as 
we could with the limited plant material that we had. We are looking at the most fundamental of these 
questions, 1 think, which is where does successful regeneration come from if we are going to plant it? So 
we decided we would plant American chestnut seedlings in a very open situation, in a clearcut or a very 
low residual basal area oak shelterwood or that we would plant under an oak canopy and treat it in such a 
way that we would have a modest increase in light to be released several years later by an overstory 
removal. Those were the two basic treatments (open setting vs. under a canopy), and again, we were 
limited by the amount of chestnut plant material we had. The other thing we wanted to do in this study 
was to compare these strategies, planting in the open versus planting under a modified stand structure. 
But also we planted on both moist sites where yellow-poplar is a competitor and on drier sites where it is 



169 



not. As Phelps pointed out, yellow-poplar can be a fairly severe competitor. In the Southern 
Appalachians. it"s maybe a little different than in Pennsylvania. But that is the nature of the study. At 
this point, all of the plantings are in eastern Kentucky, two of them are on National forest land, one at a 
college, one on a state forest, and the other on a school forest. 



We are in the process of completing the 
necessary paperwork to actually put a 
study on the Bent Creek Experimental 
Forest. It's been a long time coming, but 
we'll put in that stud\ in the next >ear or 
two. But in the current location of the 
study the treatment was a very open stand 
condition, in this case a low residual basal 
area shelterwood (Figure 3). We chose 
this particular treatment because this is 
consistent w ith the forest plan on National 
Forest Land. We would not necessarily 
have been allowed to clearcut on National 
Forest Land. We created very open 
conditions with this shelterwood cut, 
creating an open shelterwood with 20 ft" 
of residual basal area. This was also a 
midstory removal, where we took out 
perhaps 20%-30% of the basal area from 
below, using herbicide injections for that. 
Again we did those treatments on some north facing slopes, mesic sites, where \ellow-poplar was a 
competitor, as well as on south-facing slopes, where yellow-poplar was not a competitor. 




Figure 3. Open shelterwood planting site (20 ft" residual 
basal area). 



In summary, we have an open shelterwood cut on a xeric site (Figure 4), an open shelterwood cut on a 
mesic site (Figure 5). a midstory removal on a xeric site (Figure 6), and a midstory removal on a mesic 
site (Figure 7). 





Figure 4. Open shelteiAvood cut on xeric site. 



Figure 5. Open shelterwood cut on mesic site. 



170 


















^1-^, -''- 














Figure 6. Midstory removal on xeric site. 



Figure 7. Midstory removal on mesic site. 



EARLY CONCLUSIONS 

So far we don't see any differences in survival among the sites and treatments (Figure 8). Chestnut 
planted in the open grows faster than those planted under more shaded conditions (Figure 9). But all of 
the seedlings have grown. It will be another 3-4 years probably before we are really able to definitely 
assess the competitive status of chestnut planted in the open. 

This story is not complete yet. You can see the chestnut seedling in the open shelterwood site in Figure 5. 
But on the other hand, there is also yellow-poplar. We don"t know whether that chestnut can actually 
compete with yellow-poplar, the primary competitor, without supplemental competition control. We 
certainly found that oaks cannot do so. So that is really the story we are looking at here, especially in the 
southern Appalachians - whether or not the planted chestnut can keep up with some of the competition, 
notably on the higher quality sites. And especially whether it can keep up with yellow-poplar on the 
really mesic sites. It's going to be quite awhile before we are able to assess the competitive status of 
chestnut that was planted under a modified canopy and then subsequently released. That is probably 
about a decade away. 

So we do not know how much competition control is going to be necessary to establish chestnut under 
various conditions. But, 1 guess 1 do question from a logistical and economic point of view whether or 
not we can effectively reintroduce chestnut on a very, very large scale and control competition as 
intensively as may need to be done, particularly on mesic sites. It's entirely possible that we may end up 
looking at different planting strategies, one for one more xeric sites and a different strategy for mesic 
sites. I don't think we know this yet, and only time will tell, at least in the southern Appalachians. 

Finally, 1 would also like to pose a question that might spur some additional research. If we find that we 
need to adopt multiple strategies or if we simply find that planting under a modified canopy might be a 



171 



viable strategy, do we need to look at how we produce seedlings that are well adapted to the environments 
in which they are being planted? This work should really operate in tandem. I am not sure that we are 
going to get perfect answers, as we are limited in terms of the plant material that we have to v\ork with. 
But it could well be that we might want to modify the way that we produce seedlings if we are going to be 
planting them in more shaded environments. 



LITERATURE CITED 

Ayers, H. B., and W.W. Ashe. 1905. The southern Appalachian forests. U.S. Geol. Surv. Prof. Pap. 37. 
U.S. Government Printing Office, Washington, DC. 232 pp. 

Frothingham. E.H. 1917. Report on study of cut-over areas in the southern Appalachians. Unpublished 
report on file at the Bent Creek Experimental Forest, USDA Forest Service. Ashe\ ille. NC. 



172 



Steiner, K. C. and Carlson, J. E, eds. 2006. Restoration of American Chestnut To Forest Lands - Proceedings of a 
Conference and Workshop. May 4-6, 2004, The North Carohna Arboretum. Natural Resources Report 
NPS/NCR/CUE/NRR - 2006/001, National Park Service. Washington, DC. 

Ecosystem Restoration and Federal Land Policy: 
Reexamination in Light of The American Chestnut Restoration Effort 

;C7 ' Robert B. McKinstry, Jr.' 

School of Forest Resources, The Pennsylvania State University, 
University Park, PA 16802 (rbiTilO@psu.edu) 



The suite of federal laws defining the goals and policies regarding ecosystem and biodiversity 
management on federal lands begin with the presumption in natural systems that "if it ain't broke, don't 
fi.x it." These laws presume that waiting and studying the system will ( 1 ) enable us to develop a 
comprehensive plan of action that will prevent or minimize future harm and (2) waiting will not cause 
further harm. 

Beginning with the National Environmental Policy Act ("NEPA")" and continuing with the various laws 
defining the terms of ecosystem management on federal lands, those laws have required strict scrutiny of 
new actions to assure that they do not disrupt existing conditions, which are deemed ""natural." This 
presumption is reflected in the requirement under the NEPA regulations that every environmental impact 
statement include consideration of the ""no-action" alternative.^ Although the leading federal law directly 
aimed at biodiversity conservation, the Endangered Species Act,' recognizes the need for active 
intervention and management in calling for the development and implementation of recovery plans for 
threatened and endangered species, even that Act focuses heavily upon the prevention of human activities 
adversely affecting those species or their critical habitat, in fact, prohibitions on ""taking" individuals can 
often prevent application of actions to restore habitat and support the species (Bean 2003). In many 
cases, if not the majority of cases, the underlying presumption in favor of leaving ""Mother Nature" alone 
without a good showing that we will not unduly disturb her is a good approximation of reality, insofar as 
it has been human activities that have caused much of the environmental disruption that has been 
observed to date. 

However, human influence has become so pervasive globally that the presumption against human activity 
is no longer valid in many cases. Human intervention is often necessary to maintain or restore damaged 
ecosystems and species populations. In these cases, human intervention is required to manage a ""natural 
system" to preserve the system's existing characteristics or to restore the system (Flannery 2001 ). In 
these cases, the laws designed to prevent injury by requiring that the government wait and study before 
acting can also deter or prevent altogether ameliorative actions. For example, human introduction of 
pests and diseases can threaten species and disrupt ecosystems without human intervention. Introduction 
of the chestnut blight reduced a species which formerly was found throughout the hardwood forests of the 
eastern North America and comprised as much as 25 to 50% of the forest composition where it was 
dominant to a shrub growing from root sprouts (Oak 2002: Rhoades 2001; Russell 1987). Human 
introduction of the wholly adelgid threatens the hemlock. Human suppression of fire in the United States 
has so disrupted natural ecosystems based upon repetitive small fires that many areas are threatened with 



Maurice K. Goddard Professor of Forestry and Environmental Resources Conservation. I would like to thank 
Emily Lisy for her assistance in research for this article. 
- 42U.S.C. §§4321-4370f 

^ These include, inter alia, the National Forest Management Act, 16 U.S.C. §§ 1600-1614, the Federal Land Policy 
and Management Act, 43 U.S.C. §§ 1701-1785, and ithe Wilderness Act. 16 U.S.C. §§ 1131-1136. 
^ ^ff 40C.F.R. § 1504(d). 
^ 16 U.S.C. §§ 1531-1544. 

173 



wildfires that would destroy the system rather than restore it. In these cases, human action is required to 
restore or maintain the balance. 

The presumption against action, coupled with the requirement for comprehensive study and planning 
before taking action, can be particularly problematic for restoration actions. The complexity of natural 
systems and limitations on knowledge often makes it difficult or impossible, as a practical matter, to 
develop a comprehensive plan of action ah witio in cases of restoration. In natural systems, restoration 
often requires adaptive management using the process of circumscribed trial and error followed b\ 
modified trial and error. 

Without the creation and consistent application of exceptions to the no action presumption of current 
federal laws, land managers can feel that their hands are tied and they are unable to undertake the required 
restoration actions quickly. Even where exceptions are applied, land managers may act at the risk that a 
federal court will second guess their action. Because additional study entails both cost and delay, 
requirements for study before restoration can deter any ameliorative action altogether or delay it until a 
species has become threatened or endangered and weathered the storm of listing under the Endangered 
Species Act. Where other political pressures demand some action. Congress has therefore occasionally 
intervened to provide relaxed or expedited standards to allow restoration programs. For example, in 
1982, Congress created the experimental population concept in ESA to allow more flexibilit\ to 
encourage or assist programs for reintroduction of threatened or endangered species. More recently, in 
2003, Congress enacted the Healthy Forests Restoration Act to relax certain procedural requirements 
applicable to efforts to reduce excess fuel in forests using logging to prevent fires that threaten fire 
ecosystems as well as human homes (creating the political demand for some action). 

These measures have been controversial. Moreover, at best, they represent Band-Aids dealing w ith 
limited aspects of a more pervasive problem that will emerge more frequently as human populations and 
infiuences increase outside of reserved forest and parkland areas. A more systematic approach is required 
to deal with the issue of actions to restore damaged species and ecosystems on federal lands. An 
opportunity now exists to develop a more systematic approach in the context of implementing the 
recommendations contained in the NEPA Task Force Report to CEQ: Modernizing NEPA 
Implementation (CEQ 2003). That report called for measures broadening use of exceptions to the 
presumptions against action and establishment of an adaptive management work group to explore 
increased use of adaptive management techniques. The experience and challenges posed b\ American 
chestnut restoration could inform this process. This paper will, therefore, examine the problems with the 
existing federal model and possible solutions in the context of the efforts of the American Chestnut 
Foundation to restore the American chestnut as a dominant forest tree in the east. 



The American Chestnut Restoration Efforts 

The American chestnut {Castanea (kmtata [Marsh.] Borkh.) was, for the last 2000 years, a major 
component of the forests of eastern North America. Its range extended t>om central Alabama north to 
Vermont, New Hampshire and Maine and west from southern Ontario through Ohio, soutiiern Indiana. 
Kentucky, and Tennessee (Russell 1987). It was a co-dominant in the Oak-Chestnut Forest (now known 
as the Oak Hickory Forest) that formerly covered much of the northern and eastern Appalachian region 
(Kircher & Morrison 1988 at p. 58). In the heart of its range, the chestnut could comprise 25 to 50% of 



' 16U.S.C.§ l539(j);EndangeredSpecies Act Amendments of] 982, Publ. L. No. 97-304. 96 Stat. 1411. 1422. §6: 
see H.R. Report No. 97-567. at 17. 33-35. reprinleJ al 1982 U.S.C.C.A.N. 2807. 2817. 283333-2835. 
' Pub. L. 108-148, 117 Stat. 1888 (Dec. 3, 2003), c-oc////Vc/ a/ 16 U.S.C. §§ 6501-6591. 

174 



the forest cover (Oak 2002; Rhoades 2001 ; Russell 1987). It represented up to 70 percent of the wood 
volume on some slope forests. (Paillet 2002). 

As a dominant species, it played an important ecological and economic role. Its nuts were prolific and 
consistently produced. As such, they provided mast supporting many species of wildlife (Wright & 
Kirkland 1999-2000; Lord 1998-1999; Morgan & Schweitzer 1999-2000). It was a valuable source of 
lumber for furniture, construction fencing and poles, and was used in tanning (Buttrick 1915; Russell 
1987). Its nuts were also widely used and marketed for a human food, being widely used for chestnut 
stuffing (Buttrick 1915). According to the author's grandfather, its nuts were sweeter and much tastier 
than the chestnuts available today. 

In 1904 the fungus, Cryphonectria parasitica (Murrill) Barr, was introduced from Asia to the Bronx New 
York, causing the chestnut blight (Oak 2002). The blight rapidly spread throughout the range of the 
American chestnut, killing 50 to 99 percent of American chestnut trees by 1940 (Oak 2002). Today, the 
American chestnut has been virtually extirpated throughout its range as a canopy tree. 

The American chestnut persists as an understory or shrub species as a result of root sprouting. Virtually 
all trees produced from sprouting are eventually killed by the blight, although a few persist to produce 
some seeds. As a result of its sprouting behavior, the American chestnut "is in a somewhat unique 
situation among candidates for species restoration" in that "probably millions of sprouts" remain, despite 
the extirpation of the tree. As a result, the chestnut "is ranked G-4" "widespread, abundant and 
apparently secure globally" but presenting "some cause for long-term concern" (Irwin 2003). 

The disappearance of the chestnut as a canopy tree, nevertheless, has had "profound" ecological 
implications (Oak 2002). Its disappearance has been implicated in the decline of oaks, particularly in the 
southern Appalachian region (Oak 2002). Because its nut production was more prolific and reliable and 
its nuts more nutritious than its replacement, the acorn, its disappearance, like the disappearance of other 
keystone species, has likely had wide implications for many wildlife species (Lord 1998-1999). The 
decline of the Allegheny woodrat {Neotoma magister), a species that has been federally listed as 
endangered, has been attributed to the disappearance of the American chestnut (Wright & Kirkland 1999- 
2000). Numbers of eastern wild turkeys have also likely been depressed (Morgan & Schweitzer 1999- 
2000). 

The importance of the chestnut to human and natural systems has created incentives for several programs 
that hope to restore the American chestnut by creating blight lesistant strains and reintroducing these 
strains throughout the chestnut's former range. Although these programs are collaborative and involve 
the Joint efforts of individuals and private institutions, universities and state and local government 
agencies, the efforts have been privately led. 

Four restoration methods are being pursued. The first effort, which is described here at length and 
appears poised to begin implementation of the actual restoration effort, is that of the American Chestnut 
Foundation; that effort involves backcrossing American chestnuts with oriental chestnuts to incorporate 
blight resistance. A second effort, which also appears promising, is using recombinant techniques to 
attempt to incorporate blight resistance directly into the DNA of the American chestnut. The American 
Chestnut Cooperators' Foundation is involved in a program of breeding the more blight resistant 
American chestnuts that can be found in the wild. Other efforts seek to utilize viruses that cause the 
blight fungus to become hypovirulent. 

The American Chestnut Foundation was founded in 1984 to pursue the backcross restoration method. 
The Foundation's efforts have now approached the point where reintroduction and restoration of 
American chestnut appears feasible in the near future. The Foundation has established a program which 



75 



involved crossing the American chestnut with the Chinese chestnut and then backcrossing the progeny 
(Fi) with American chestnuts found flowering in the wild three times. After each backcross. the 
Foundation selected only the trees with blight characteristics and the phenotypic characteristics of the 
American Chestnut. This has produces trees that are genetically 1 5/16 American and blight resistant. 
The Fi- B3 generation is then intercrossed two times, again selecting for blight resistance and the 
American chestnut phenotype, to produce offspring that are homozygous for resistance and otherwise 
include predominantly American chestnut genes. The objective of this program is to produce "backcross 
trees [that] will fall within the range of American chestnut ta.xonomic characteristics as understood from 
monographs and voucher specimens, [although] known to carrv alleles from Chinese chestnut. [Its h\brid 
origin will not be recognizable except for its] blight resistance and on a DNA level." This breeding 
program has now produced the B3-F2 trees for seed orchards that will produce the B3-F3 nuts that will be 
planted in the forests as part of the final restoration effort, described below (Hebard 2002. 2003; Irwin 
2003; Burnham 1991). 

The next step in this effort will entail reintroducing American chestnut into the forest by planting B3-F3 
seedlings from the seed orchards. How this will proceed has not been settled, but it will involve a series 
of steps employing adaptive management techniques. It will likely first entail installing experimental 
small plantations on existing openings on federal, state and other public or institutional lands. These 
plantations and any impacts will be monitored. Moreover, since the ecology and needs of the American 
chestnut are not fully known, a certain amount of experimentation will be needed to detemiine optimum 
planting techniques, favorable soil characteristics and the need for management techniques, including 
possibly controlled burning (Klinger 2000; Perry 2003) The American chestnut is released in sunlight, so 
the creation of clearings in forests will likely be required. After the plantations prove successful, the 
reintroduction will occur on a larger scale. 

Resistant chestnuts created by other methods for will also likely be poised for use in reintroduction in the 
near future. Genetically engineered American chestnuts containing genes that ma\ create blight 
resistance have also been created by the second method and may be ready for introduction by the end of 
the decade. Inoculation of American chestnuts with the hypovirulent virus may also proceed in federal 
lands. 

file introduction of the backcrossed American chestnuts and other resistant strains can proceed without 
procedural impediments on state, institutional and prixate lands. However, it may run afoul of the federal 
presumption against action when attempted on federal lands. As discussed more fully below, existing law 
provides mechanisms that ought to allow the flexibility to allow these efforts to proceed without 
significant delay or cost. However, third part>' litigants who are concerned about possible adverse effects, 
the courts, and land managers who are concerned about the threat of litigation or othenvise unwilling to 
depart from standard operating procedures ina\ impose costly environmental impact stud\ procedures that 
could slow reintroduction efforts on federal lands or make some infeasible. A consistent federal policy 
designed to reverse the presumption against taking action in cases of reintroduction and restorations could 
help avoid these costG and delays and better serve the underlying intent of NEPA and other federal laws 
intended to incorporate environmental concerns into federal decision making. 



Wi 1 AT IS Needed in a Restoration Efeort 

In assessing how federal laws might be applied to restoration efforts and what form a federal restoration 
policy might take, it is helpful Hrst to consider the characteristics of an cffccti\c restoration project. 
These characteristics have described by a variety of authorities in a variet> of contexts (Adams cl al. 
1998; Frelich & Puettman 1999; Marker c/ c;/. 1999; Henry & Lucash 2000-200 1 ; Gjerstad, D. 2000- 



176 



2001 ) and can be applicable to efforts to restore the American chestnut (Craddock, J. H. 2000-2001; Irwin 
2003). 

All restoratioii programs require affirmative actions to reintroduce a species, to restore habitat or site 
conditions, and to manage the species and site after introduction both to maintain the species or system 
that has been restored and to make any changes found to be necessary. Physical site modification and 
planting or release of native or formerly native species will be required and physical management to 
maintain or restore soil conditions or control plants or animals following introduction will usually be 
required. 

In most cases, reintroductions must be initiated with limited knowledge of the ecology, threats and 
requirements of the species or system to be restored. This requires that any restoration be preceded by 
study of the historic or paleohistoric records. The American Chestnut Foundation has been gathering this 
information since its inception. However, much of the information is simply unavailable and cannot be 
obtained without actual experience in the field. 

Adaptive management techniques will be required throughout the reintroduction effort. Reintroduction, 
itself will require at least two steps. The initial step of limited reintroduction or restoration will often 
involve planting test plots or releasing a limited number of individuals in somewhat controlled conditions 
and monitoring these areas. The American Chestnut Foundation has developed seed orchards for its 
backcrosses using several strains of American chestnut, with the American chestnut genetic material 
being gathered from a variety of trees in several different regions. Trees from seed orchards containing 
the ''local" genetic material will be planted in these test plots. In this step, infomiation on techniques for 
restoration, the needs of the species or system being restored, management techniques, possible threats to 
the restoration and impacts, if any, of the restoration can be identified. The techniques and management 
can then be adapted to structure the actual reintroduction. 

The actual reintroduction will involve planting the American chestnuts at sites throughout each relevant 
region to encourage the widespread introduction. This step will involve site selection, site preparation 
and planting guidelines based on the experience in the test plots. The reintroduced species will require 
monitoring and will likely require management. For example, fire likely played an important role in 
chestnut ecology, such that controlled burns may be required for site preparation and management (Perry 
2003). Use of herbicides to control competing vegetation may also benefit American chestnut restoration. 
Monitoring will be required to determine issues such both the success of the reintroduction, and the 
genetic and phenotypic characteristics of new trees, the impacts of the reintroduction and possible 
modifications of management techniques. Corrective actions will likely often be required. 



Legal and Practical Barriers to Reintroduction on Federal Lands 

Although the basic structure of federal environmental laws and regulations governing the use of our 
public lands begins with the presumption that new actions should be deferred and studied pending 
implementation, these laws and regulations provide sufficient flexibility to structure a program that will 
allow restoration programs such as that proposed for the American chestnut to be readily implemented 
without excessive cost or delay. In many cases, however, the opportunity to use this flexibility is 
squandered due to the unwillingness of federal managers or regulators to take advantage of these 
opportunities, whether due to fear of making a mistake, an unwillingness to depart from standard 
operating procedures, or concern regarding possible litigation. Flexibility can also be hampered by 
litigation brought by groups equally unwilling to depart from standard operating procedure, often due to 
suspicion regarding the motivation of the federal managers or, more frequently, their political superiors. 



177 



Courts, too, contribute to this confusion, upholding actions to expedite restoration activities in some cases 
and halting these activities to require additional costly process in virtually identical situations. 

These conflicting results could be avoided with the formal adoption of a consistent federal policy . _" 

applicable across all agencies towards the treatment of species or ecosystem restoration plans under ^ 
federal law. Sufficient flexibility likely exists within the existing statutory framework to allow such a 
policy to be implemented by regulation. Executive Order or as formal guidance. The new federal policy 
should explicitly provide flexibility and encourage immediate implementation of restoration actions 
employing adaptive management under existing law. regulation and guidance. Adoption of a consistent, 
interagency, written policy would have several advantages. It would limit the discretion of the federal 
officials unwilling to take a risk or depart from standard operating procedure, it would provide written 
direction to courts that would help avoid the t>pes of conflicting results that t\pify the current legal :_-, 
landscape and support the concerns of the federal officials whose inaction so often stymies proactive 
restoration. Finally, if it includes adequate safeguards against abuse, it might serve to alle\ iate the 
concerns among the groups who bring the litigation. Care must be taken in crafting such a polic\ to 
assure that it is, not, in fact, subject to abuses, but too much care in that regard could eviscerate the intent 
of the policy. The outline and justification for such a policy, using the American chestnut restoration as a 
model, are provided here. 

Several environmental laws are applicable or potentially applicable to implementation of the American 
chestnut restoration on federal lands or similar efforts. As the fundamental environmental law governing 
all government planning and any federal action, NEPA^ can apply and will be the focus of this article. 
The National Forest Management Act ("NFMA"),'' although potentially applicable on National Forest 
lands used in the effort, will likely apply only in the context of NEPA. The Endangered Species Act 
("ESA") '" although inapplicable to the American chestnut restoration, is the federal law most applicable 
to restoration efforts and will be discussed because of the experience that has been gained under ESA 
regarding some of the difficulties in applying reforms encouraging voluntary action. A variety of federal 
laws are potentially applicable to the introduction of genetically engineered blight resistant American 
chestnuts, depending upon the genes introduced. The Federal Insecticide. Fungicide and Rodenticide Act 
C'FIFRA") is potentially applicable to the efforts to use of the virus to induce h\povirulence in the 
chestnut blight fungus. The laws governing genetically modified organisms ("GMOs") and pesticide 
regulation are beyond the scope of this article. 

NEPA - Establishing the Structure of Environmental Decision-Making Governinu Restoration. 

The National Environmental Policy Act has often been described as the foundation or the cornerstone of 
modern American environmental law. Enacted in 1970. it was the first major federal en\ ironmental law 
enacted in the "environmental decade" and has profoundly affected the development of en\ ironmental 
policy. It established: 

the continuing policy of the Federal Government ... to use all practicable means and measures 
... in a manner calculated ... to create and maintain conditions under which man and nature can 
exist in productive harmony. . .'' 



" 42U.S.C.§§4321-4370f. 
" 16U.S.C.§§ 1600-1614. 
'•• 16U.S.C. §"§ 1531-1544. 
" 7U.S.C.§§ 136-136y. 
'- 42 U.S.C.§ 4331(a). 



78 



NEPA's basic structure is simple. It seeks to require all government agencies to incorporate 
environmental consideration into all aspects of their planning, using an interdisciplinary approach. It 
seeks to do this through two mechanisms. First, it requires that each federal agency to use an 
interdisciplinar>' approach to incorporate environmental considerations into its planning and "include in 
every recommendation or report on proposals for legislation and other major Federal actions significantly 
affecting the quality of the human environment, a detailed statement by the responsible official on," inter 
alia, the environmental impacts, unavoidable adverse environmental effects, and alternatives to the 
proposed actions.'" The United States Supreme Court has noted: 

NEPA has twin aims. First, it "places upon an agency the obligation to consider every significant 
aspect of the environmental impact of a proposed action." [citation omitted] Second, it ensures 
that the agency will inform the public that it has indeed considered environmental concerns in its 
decisionmaking process, [citation omitted]. Congress in enacting NEPA, however, did not 
require agencies to elevate environmental concerns over other appropriate considerations, 
[citation omitted].'"' 

NEPA created the President's Council on Environmental Quality ("CEQ") to serve as an independent 
group overseeing all programs to assure consideration of environmental impacts of all federal policies. 
CEQ was given the authority oversee the environmental impact assessment requirements by promulgating 
rules governing agency implementation, overseeing agency implementation and resolving disputes. 
CEQ's regulations, thus, govern the structure under which restoration projects must be reviewed and 
assessed.'^ 

CEQ regulations provide that agencies should "[i]nterpret and administer the policies, regulations, and 
public laws of the United States in accordance with the policies set forth in the Act and in these 
regulations." Each agency must also adopt its own "procedures" incorporating the CEQ's requirements 
and supplementing those regulations as necessary.'^ The regulations call for "[i]ntegrat[ing] the 
requirements of NEPA with other planning and environmental review procedures required by law or by 
agency practice so that all such procedures run concurrently rather than consecutively." 

The central element of the NEPA process is the preparation of an environmental impact statement, 
assessing impacts across a range of media and concerns, developing alternatives, identifying mitigating 
measures and identifying unavoidable impacts.''' Although the scope of the "statement" required by law 
is unspecified, in application, preparation of an EIS has become an expensive and time-consuming 
venture, involving analysis of multiple potential impacts and development of multiple alternatives. 
Moreover, these costly ventures often result in delay, since CEQ regulations mandate that an agency defer 
actions that will adversely affect the environment or limit alternative choices while the process is 
unfolding."*^ 

Moreover, the EIS process, as applied, involves a top-down, comprehensive planning process that 
assumes that knowledge of impacts and effects of alternatives either exists or can be gathered before 



'M2U.S.C. §4332. 

'^ Baltimore Gas & Electric Co. v. Natural Resources Defense Council, Inc., 462 U.S. 87, 97, 103 S.Ct. 2246, 2252 

(1983). 

'^ These regulations appear at 40 C.F.R. Parts 1 500- 1517. 

'^ 40 C.F.R. § 1500.2(a). 

" 40 C.F.R. §§ 1505.1, 1507.3. 

'^ 40 C.F.R. § 1500.2(c). 

'^ 40 C.F.R. §§ 1501.7 (scoping); 1502.1-1502.25 (requirements for EIS preparation); 1503.1-1503.4 (public 

comment and response); 1505.2 (preparation of record of decision). 

-" 40 C.F.R. § 1506.1. 

179 



implementation. This assumption is invalid in many restoration projects, where use of adaptive 
management techniques may better serve the underlying statutory intent of environmental protection. 

NEPA is intended to be a tool to assure that environmental concerns are incorporated into decision- 
making, not a mechanism to slow down and make actions more costly. CEQ has. therefore. b\ regulation, 
established several mechanisms to expedite the NEPA process, and federal agencies have incorporated 
these elements into their regulations, guidance and procedures. Thus, before undertaking an EIS, in many 
cases, agencies will undertake a less costly and less time consuming "mini-EIS" known an environmental 
assessment. This is described by one court considering a private proposal involving a mechanism 
reducing conflicts between private development and efforts to restore endangered species, as follows: 

NEPA is not designed to prevent all possible harm to the environment: it foresees that ^^ 
decisionmakers may choose to inflict such harm, for perfectly good reasons. Rather. NEPA is 
designed to influence the decisionmaking process; its aim is to make government officials notice 
environmental considerations and take them into account. By regulation, an agenc\ considering 
whether an action would require preparation of an EIS must prepare a brief, preliminary 
evaluation, called an environmental assessment ("EA")."' 

The CEQ regulations also require all federal agencies to designate categories of actions that "do not 
nomially require either an environmental impact statement or an environmental assessment."" A number 
of these "categorical exemptions" appear to apply to the American chestnut restoration and could enable 
these efforts, despite the overall presumption in favor of inaction. 

Agency Procedures Relevant to Restoration: NEPA Compliance and Categorical Exclusion as Applied to 
the American Chestnut Restoration Efforts on Federal Lands 

Although many federal agencies own and manage lands that may be involved in restoration efforts, most 
federal lands that might be involved in restoration efforts for the American chestnut will be managed by 
either the National Park Service ("NPS") or the United States Forest Service."^ AccordingK . the impact 
of NEPA and the federal presumption against action must be examined in the context of the regulations 
and guidelines of those two agencies. These regulations and guidelines provide sufficient flexibilit> to 
authorize the initiation of restoration employing adaptive management avoiding the presumption against 
initiating actions without delay. However, as will be discussed further below, they do not provide 
sufficiently clear or consistent guidance to avoid the risk that these efforts could be derailed by public 
opposition, concerns of individual agency personnel or litigation. 



- ' ( 'enter for Blologleal Diversity v. United States Fish <& midlife Sen-ice, 202 P. Supp. 2d 594. 647 (W.D. Tex. 

2002);.vt'L'40C.F.R. §§ 1501.4(b). 

" 40C.F.R. §§ 1507.3(b){2)(ii); 1508.4. 

'^ Most other significant federal lands fall under the jurisdiction the Bureau of Land Management (BLM"). 

However, lands under BIM jurisdiction arc found primariK in the West, outside of the range of the American 

chestnut. The Department of Defense also manages extensive lands held as militaiy bases, training areas and target 

and bombing ranges. DOD has developed a program for ecosystem management for its lands with the assistance of 

The Nature Conservancy (Leslie et al. 1996). However, DOD has not yet been actively engaged in the American 

chestnut restoration project. Moreover. DOD has a different mission than land management agencies and faces less 

significant budget constraints than do other land managers, such that it can readily retain outside contractors, such 

that delay and cost is less likely to deter restorative action. 

180 



N PS Procedures and Categorical Exclusions: The Management Policies of the National Park Service 
incorporate a variety of policies that encourage ecosystem and species restoration and would facilitate the 
reintroduction of the American chestnut on Park Service Lands (US Dep't of Interior 2000). These 
policies appear to reverse the general presumption against action for restoration and could serve as the 
kernel of a more general model. 

The Park Service policies encouraging restoration arise from the mandate contained in the National Park 
Service Organic Act that the Service 

promote and regulate the use of . . . national parks, monuments, and reservations. . . by such 
means and measures as confomi to the fundamental purpose . . .to conserve the scenery and the 
natural and historic objects and the wild life therein . . .and by such means as will leave them 
unimpaired for the enjoyment of future generations."'* 

The Park Service has broadly interpreted the non-impairment mandate to include any impact that would 
impair resources including past and external impacts and to require monitoring and affirmative action to 
address those impacts.'' The Service's prohibition against intervention in natural or physical processes 
excludes efforts "[t]o restore natural ecosystem functioning that has been disrupted by past or ongoing 
human activities;" and the Service has adopted a restoration policy mandating action unless otherwise 
directed: 

The Service will re-establish natural functions and processes in human-disturbed components of 
natural systems in parks unless otherwise directed by Congress ...Impacts to natural systems 
resulting from human disturbances include the introduction of exotic species; the contamination 
of air, water, and soil; changes to hydrologic patterns and sediment transport; the acceleration of 
erosion and sedimentation; and the disruption of natural processes. The Service will seek to 
return human-disturbed areas to the natural conditions and processes characteristic of the 
ecological zone in which the damaged resources are situated. The Service will use the best 
available technology, within available resources, to restore the biological and physical 
components of these systems, accelerating both their recovery and the recovery of landscape and 
biological-community structure and function. Efforts may include, for example: 

• Removal of exotic species; 

• Removal of contaminants and non-historic .structures or facilities; 

• Restoration of abandoned mineral lands, abandoned or unauthorized roads, areas over- 
grazed by domestic animals, or disrupted natural waterways and/or shoreline processes; 

• Restoration of native plants and animals.'^ 

The Service's Policies further encourage establishment of public-public and public-private partnerships to 
accomplish these goals." 



-' 16U.S.C. § 1. 

"^ U.S. Department of the Interior, National Park Service, Management Policies 2001 {"NPS 2001 Policies'"). 

NPSD1416, §§ 1.4.5, 1.4.7, 1.5(2000). The Park Service is in the process of updating its Management Policies, 70 

Fed. Reg. 60852 (October 19, 2005), see U.S. Department of the Interior, National Park Service, Draft 2006 NPS 

Management Policies {'"Draft NPS 2006 Policies"), found at 

http://parkplanning.nps.gov/document.cfm?projectId=13746&documentID=12825 (last visited October 20, 2005), 

§§ 1.4.4- 1.4.7. 

"*' NPS 2001 Policies, supra. §§4.1.5,4.1; see also. id. § 4.4. 1 . These policies remain substantially unchanged 

under the proposed revisions to the policies. See Draft NPS 2006 Policies, supra, §§4.1.5,4.1; see also. id. § 4.4. 1 . 

-' NPS 2001 Policies, supra. §§ 4. 1 .4, 4.2, Draft NPS 2006 Policies, supra. §§4.1 .4, 4.2. 

181 



The NPS management principles broadly encourage restoration of natural populations, genetic diversity 
for those population and natural system and include specific standards governing restoration of extirpated 
species."** Fire management is encouraged and integrated pest management is allowed to protect rare, 
threatened and endangered or unique populations. I he guidelines" requirement to restore native species 
and to remove or exclude exotic species creates some ambiguity for restoration of the American chestnut 
using the back-cross or genetic engineering methods, in that the Chinese chestnut genes might be 
considered exotic. However, the facts that, even in the back-crosses, the exotic genetic materials will 
constitute less than 10% of the genes and the plants will phenotypically be American chestnut, suggest 
that the new species should better be considered native. Moreover, the guidelines specifically pro\ ide 
that an exotic species may be introduced where all measures have been taken to minimize harm and the 
species is "[a] closely related race, subspecies, or hybrid of an extirpated native species; or [a]n improved 
variety of a native species in situations in which the natural variety cannot survive current, human-altered 
environmental conditions; or [u]sed to control another. alread\ -established exotic species."" These 
exceptions, coupled with the mandate to restore systems and species would seem to encompass all of the 
methods for chestnut restoration. 

The Park Service has incorporated its presumption in favor of restoration activity into its NEPA policies. 
However, these policies are not as well developed as its management guidelines. The Park Serv ice 
categorically excludes a variety of actions related to restoration projects from requirements for an EIS or 
EA. These include: "[d]esignation of environmental study areas and research natural areas." 
"[s]tabilization by planting native plant species in disturbed areas," ''[rjestoration of noncontroversial 
native species into suitable habitats within their historic range and elimination of exotic species," 
"removal of park resident individuals of non-threatened/endangered species which pose a danger to 
visitors, threaten park resources or become a nuisance in areas surrounding a park, when such removal is 
included in an approved resource management plan", and grant programs related to these activities.' 

Forest Service Procedures and Categorical Exclusions: The Forest Service lacks a comprehensive 
restoration policy. The Service has a native plants policy, favoring use of native plants. The Service has 
also established the following categorical exclusion from NEPA documentation for planting native 
species: 

Regeneration of an area to native tree species, including site preparation which does not involve 
the use of herbicides or result in vegetation type conversion. Examples include but are not limited 
to: a) Planting seedlings of superior trees in a progeny test site to evaluate genetic worth and b) 
Planting trees or mechanical seed dispersal of native tree species following a fire. Hood, or 
landslide. ' 



28 



NPS 2001 Policies, supra. §§ 4.4. 1 . 4.4. 1 .2, 4.4.2.2 (see discussion infra): Draft NPS 2006 Policies, supra. § 
4.4.1.4.4.1.2.4.4.2.2. 

A'AS' 2001 Policies, supra. §§ 4.4.4. The new guidelines allow introduction of "exotic" species where "when all 
feasible and prudent measures to minimi/e the risk of harm have been taken, and 

• it is a closely related race, subspecies, or hybrid of an extirpated native species, or 

• it is an improved variety of a native species in situations in which the natural variety cannot survive current, 
human-altered environmental conditions. . . Draft NPS 2006 Policies, supra. § 4.4.4. 

United Slates Depailment ofthe interior. Department of Interior Manual. National Environmental Polic\ Act 
implementing Procedures for the National Park Service. 516 DM 6. App.7. § 7.4E(3).(4).(6X7). F(i). found at 
http://elips.doi.gov/elips/release/351 1 .htm . 

■ ' Forest Service Handbook ("FSH") 1 909. 1 5 - Environmental Policy and Procedures Handbook ^ 3 1 .2(5) , found at 
http://\vvvvv.fs.fed.us/emc/nepa/includes/epp.htm/^c3l . By regulation, the Forest Ser\ ice designates its NEPA 
procedures in the Forest Service Handbook. 36 C.F.R. § 200.4. 

182 



The NPS management principles broadly encourage restoration of natural populations, genetic diversity 
for those population and natural system and include specific standards governing restoration of extirpated 
species." Fire management is encouraged and integrated pest management is allowed to protect rare, 
threatened and endangered or unique populations. The guidelines' requirement to restore native species 
and to remove or exclude exotic species creates some ambiguity for restoration of the American chestnut 
using the back-cross or genetic engineering methods, in that the Chinese chestnut genes might be 
considered exotic. However, the facts that, even in the back-crosses, the exotic genetic materials will 
constitute less than 10% of the genes and the plants will phenotypically be American chestnut, suggest 
that the new species should better be considered native. Moreover, the guidelines specifically provide 
that an exotic species may be introduced where all measures have been taken to minimize harm and the 
species is "[a] closely related race, subspecies, or hybrid of an extirpated native species; or [a]n improved 
variety of a native species in situations in which the natural variety cannot survive current, human-altered 
environmental conditions; or [u]sed to control another, already-established exotic species."" These 
exceptions, coupled with the mandate to restore systems and species would seem to encompass all of the 
methods for chestnut restoration. 

The Park Service has incorporated its presumption in favor of restoration activity into its NEPA policies. 
However, these policies are not as well developed as its management guidelines. The Park Service 
categorically excludes a variety of actions related to restoration projects from requirements for an EIS or 
EA. These include: "[djesignation of environmental study areas and research natural areas," 
"■[sjtabilization by planting native plant species in disturbed areas," "[r]estoration of noncontroversial 
native species into suitable habitats within their historic range and elimination of exotic species," 
"removal of park resident individuals of non-threatened/endangered species which pose a danger to 
visitors, threaten park resources or become a nuisance in areas surrounding a park, when such removal is 
included in an approved resource management plan", and grant programs related to these activities."^*' 

Forest Service Procedures and Categorical Exclusions: The Forest Service lacks a comprehensive 
restoration policy. The Service has a native plants policy, favoring use of native plants. The Service has 
also established the following categorical exclusion from NEPA documentation for planting native 
species: 

Regeneration of an area to native tree species, including site preparation which does not involve 
the use of herbicides or result in vegetation type conversion. Examples include but are not limited 
to: a) Planting seedlings of superior trees in a progeny test site to evaluate genetic worth and b) 
Planting trees or mechanical seed dispersal of native tree species following a fire, flood, or 
landslide.^' 



-** NPS 2001 Policies, supra. §§ 4.4. 1 , 4.4. 1 .2, 4.4.2.2 {see discussion, infra); Draft NPS 2006 Policies, supra, § 

4.4.1,4.4.1.2,4.4.2.2. 

"^ NPS 2001 Policies, supra. §§ 4.4.4. The new guidelines allow introduction of "exotic" species where "when all 

feasible and prudent measures to minimize the risk of hann have been taken, and 

• it is a closely related race, subspecies, or hybrid of an extirpated native species, or 

• it is an improved variety of a native species in situations in which the natural variety cannot survive current, 
human-altered environmental conditions. . . Draft NPS 2006 Policies, supra, § 4.4.4. 

United States Department of the Interior, Department of Interior Manual, National Environmental Policy Act 
Implementing Procedures for the National Park Service, 516 DM 6, App.7, § 7.4E(3),(4),(6X7), F(I), found at 
http://elips.doi.gov/elips/reIease/35 1 1 .htm . 

'' Forest Service Handbook ("FSH") 1909.15 - Environmental Policy and Procedures Handbook ^ 31.2(5) , found at 
http://www.fs. fed. us/emc/nepa/includes/epp.htm#c3 1 . By regulation, the Forest Service designates its NEPA 
procedures in the Forest Service Handbook. 36 C.F.R. § 200.4. 

182 



any other notice used to inform interested and affected persons of the decision to proceed with or 
to implement an action that has been categorically excluded/' 

Alternative Federal Restoration Models 



There are examples of alternative federal models that expressly apply to restoration actions and mandate 
action that has been categorically excluded from the predominant presumption against action. Restoration 
actions are authorized and even required without the NEPA procedures in the case of endangered and 
threatened species under the federal Endangered Species Act ('■ESA")36 and for contaminated sites and 
spills under the Comprehensive Environmental Liability and Compensation Act (■■CERCLA"").37 
Although these programs provide some examples of mechanisms whereby the delays and costs incident to 
the predominant model might be avoided while preventing abuses and still incorporating environmental 
planning and public participation into federal programs, they also provide examples of some of the pitfalls 
in tiy ing to balance these considerations. In some cases, in trving to strike this balance, these programs, 
particularly the CERCLA program, have generated even greater costs and delays for some restoration 
programs. Both programs provide examples of the problems incident to the top down, comprehensive 
planning approach that predominates under the existing federal model. Finally, both pro\ ide examples of 
how even a well structured program can be undercut by the tendency of agency personnel or the courts to 
apply the standard model, notwithstanding contrary' directions. 

The Restoration Model under the Endangered Species Act: The federal Endangered Species Act 
("ESA")"^^ and the NEPA procedures and reforms developed by the United States Fish & Wildlife Service 
("FWS")^'' provides an alternative to the predominant NEPA model, but also presents examples of some 
of the pitfalls of an alternative model. Under this ESA model, restoration projects for threatened or 
endangered species or their habitat is mandated and delays can be avoided. However, in the absence of 
clear guidance or policy, the policies to encourage rapid and flexible action have been hampered b> 
unwillingness of individual agency personnel to depart from the traditional model, public controversy and 
inconsistent court rulings. Moreover, the "top down" and comprehensive planning approach envisioned 
under ESA makes its application to the American chestnut restoration and many other restoration 
programs problematic, at best. 

ESA uses the "fine filter" approach to biodiversity conservation, seeking to protect biodiversit> b\ 
protecting nationally endangered and threatened species and their habitat, with the stated purpose of 
conserving "the ecosystems upon which endangered species and threatened species depend." ^ ESA 
accomplishes this goal both by limiting actions and requiring affinnative restoration. Thus. ESA protects 
the individuals within threatened and endangered species directly, through its prohibition against 
"taking" and requires the development and implementations of recoverv plans for the species. 
Prohibited "takings" include activities which have an "incidental" adverse effect on a threatened or 
endangered species. However, the Act authorizes issuance incidental take permits allowing a property 
owner to conduct otherwise lawful activities in the presence of listed species, but requires each non- 
federal entity to deve'op an Habitat Conservation Plan calling for aft1rmati\e action to conser\'e the 
species or habitat. ESA also seeks to protect the land constituting critical habitat for endangered species. 



35 
yh 
■SI 
38 
39 
40 



/c/., 1131.2. 

16U.S.C.§§ 153I-L544. 

42 U.S.C.§§ 9601-9675. 

16U.S.C.§§ 1531-1544. 

The Fish and Wildlife Service and the National Marine Fisheries Service administer ESA. 

Id § 1531(b). 

Id. § 1538. 



184 



To that end: ( 1 ) ESA requires that critical habitat be designated;^" (2) the Act requires that each federal 
agency aid in the conservation of endangered species/"* and assure that programs that it administers, 
including grant, pennit, construction and management programs, will not jeopardize the continued 
existence of threatened or endangered species or "result in the destruction or adverse modification of 
their critical habitat;^^ and (3) ESA requires both the United States Forest Service and the United States 
Department of the Interior ("DOI") to develop a broader affirmative program to conserve "fish, wildlife, 
and plants including those which are endangered species or threatened species" and authorizes those 
agencies to acquire land as a part of that program/^ 

ESA's affirmative obligations to develop and implement recovery plans and to conserve endangered and 
threatened species and their habitat represents a departure for certain restoration programs from the 
dominant federal presumption against taking affirmative action. Moreover, FWS has adopted a series of 
categorical exclusions which, on their face, would appear to allow a policy of pennitting restoration 
activities to proceed without the delays and costs incident to development of either an EIS or EA. FWS' 
NEPA Guidance provides categorical exclusions for "[t]he reintroduction or supplementation (e.g., 
stocking) of native, formerly native, or established species into suitable habitat within their historic or 
established range, where no or negligible environmental disturbances are anticipated.''^*' Similarly, FWS 
provides categorical exclusions for "restoration of wetland, riparian, instream, or native habitats, which 
result in no or only minor changes in the use of the Affected local area.'" Prescribed burning "for habitat 
improvement purposes" when carried out consistent with the law, tire management activities, is 
categorically excluded. ^^ ESA and other FWS permit actions are excluded from NEPA requirements 
"when such permits cause no or negligible environmental disturbance." Incidental take permits that 
"individually or cumulatively, have minor or negligible effect on species covered in the habitat 
conservation plans as also excluded."*^ The issuance of recovery plans is excluded,^"* but any habitat 
conservation plan nonnally requires an EA under these procedures. ^'^ 

Notwithstanding these departures from the predominant "wait and study" federal model, the structure of 
the ESA program suffers from a number of limitations, some of which have been addressed by recent 
refomis and some of which persist despite those reforms. ESA's "fine filter" approach leaves significant 
gaps in biodiversity conservation. It only addresses threatened and endangered species and does not 
address many important biodiversity features that require restoration. The frequency of the American 
chestnut's occurrence makes it less likely that it could be listed as threatened and endangered, despite the 
ecological importance of restoration of this keystone species as a canopy tree producing nuts. Moreover, 
ESA still envisions a top-down, comprehensive planning process that is inconsistent with the needs of 
many restoration projects. The listing process, requiring "besv available science," is a prolonged and 
often contentious process that entails substantial delays. This has resulted in substantial backlogs that 
keep even those species eligible for listing off of the list for long periods of time. In theory, although not 
in practice, designation of critical habitat and recovery plans must be developed at the beginning of the 
process, when sufficient information is often unavailable, rather than through an adaptive management 



'- W§i533(a)(3). 

'' M§ 1537(a)(1). 

'' Id § 1537(a)(2); Tennessee Valley Authority v. Hill, 437 U.S. 153 (1978). 

^^ 16U.S.C. § 1534. 

*^ Department of Interior Manual, National Environmental Policy Act Implementing Procedures for the Fish and 

Wildlife Services, 516 D.M. 6, Appendix 1, § 1.4 B(6), 62 Fed. Reg. 2375 (January 16, 1997), also found at 

hUp://elips.doi.gov/elips/release/35 1 1 .htm .. 

'' M,516D.M. 6, Appendix I, §§ 1.4 B(3), (4), (5). 

^^ M, 516 D.M. 6, Appendix 1,§§ 1.4C(1), (2). 

'" M. 516 D.M. 6. Appendix 1, § 1.4 D. 

'" M,516D.M. 6, Appendix 1,§ 1.5. 

185 



process. Finally, ESA suffers from lack of flexibility, particularly with respect to the prohibition against 
takings, such that listing of a species would impair property uses and would create disincentives for 
private conservation efforts (Bean 2003; Taylor 2002). The flat prohibition against takings and 
requirements for incidental take permits, with the possible requirement for an EIS for such permits, also 
deters habitat restoration activities that would benefit the species in the long run but might incidentally 
"take" individuals. 

Some of these problems were resolved by a number of reforms initiated under DOE Secretar> Bruce 
Babbitt in the Clinton Administration (Bean 2003; Taylor 2002), some of which might facilitate the 
American chestnut restoration, could they be applied. The Candidate Conserxation program seeks to 
encourage proactive programs to eliminate the necessity for regulator) controls by encouraging private 
conservation activity directed to unlisted species that would qualify for listing to reduce the threats to 
such declining species, and thus avoid listing. The requirements and procedures are incorporated into 
Candidate Conservation Agreements ("CCAs"); CCAs assure non-federal landowners that they can 
continue agreed-upon activities even if the species becomes listed in the future, and thereby a\oid 
regulatory controls (Bean 2003; Ruhl 2004).^' The program does not, however, apply to or encourage 
similar activities by federal landowners. Safe Harbor Agreements encourage voluntary actions by 
landowners to protect endangered species, in return for protection against future changes (Bean 2003; 
Taylor 2002; Ruhl 2004).^" FWS has recently proposed amendments to its permitting rules that, if 
adopted, would provide greater flexibility for habitat enhancement activities both in these programs and 
enhancement programs on federal lands, allowing incidental takes incident to programs that enhance 
habitat.^-^ 

The Candidate Conservation with Assurances mechanism, coupled with the categorical exclusion of 
habitat conservation activities might be an ideal mechanism to encourage the restoration of the American 
chestnut, which might be in danger of becoming threatened. However, the policy does not appK to 
federal lands. Moreover, problems have emerged in the application of refomis. Whether because of 
unwillingness to depart from the traditional model for federal action or concern regarding judicial review, 
federal managers have often been unwilling to apply these models. Rather than expedite implementation. 
FWS has often delayed implementation with long review, excessive requirements in the agreements, and 
insistence on studies regarding impacts and effects. At times, reforms have been slowed b\ insistence on 
a showing that individuals not be taken, despite a clear expectation that the species would benefit from 
habitat improvement. This unwillingness to depart from business as usual has led some parties to 
abandon projects and others to indicate an unwillingness to use the refomi mechanisms in the future 
(Bean 2003). Moreover, at times, the concerns of the managers have been confirmed b\ inconsistent 
results in the courts. Compare Center for Biological Diversity- v. United States Fish and Wildlife Service. 
202 F. Supp. 2d 594 (W. D. Tex. 2002) (upholding incidental take permit where pemiittec protected off- 
site conservation areas providing superior habitat for endangered species and rejecting claim that 
alternative reducing size of on-site disturbance should have been more fully developed and required) with. 
Gerber v. Norton, 294 F.3d 1 73 (D.C. Cir. 2002) (overturning incidental take permit where permittee 
protected off-site conservation area providing superior habitat, with court relying on failure to provide 
adequate comment and failure to adopt on-site plan that would minimize area of disturbance). 



United States Department of the Interior, Announcement of Final Polic> lor Candidate Conservation Agreements 
with Assurances, 64 Fed. Reg.32,726 (.lune 17. 1999). 

^' United States Department of the Interior. Announcement of Final Safe Harbor Policy, 64 Fed. Reg. 32. 717 (June 
17. 1999). 

United States Fish & Wildlife Service, Proposed Revisions to the Regulations Applicable to i'cnnits Issued 
Under the Endangered Species Act. 68 Fed. Reg. 53327 (May 3. 2003).^ 

186 



The Restoration Model under the Comprehensive Environmental Liability and Compensation Act: The 
approach adopted by the United States Environmental Protection Agency ("EPA") in pursuing hazardous 
substances remediation under the Comprehensive Environmental Liability and Compensation Act 
("CERCLA")'^ presents a useful model in two senses. The EPA model, as spelled out in the statute and 
the National Oil and Hazardous Substances Pollution Contingency Plan ("NCP"')^^ provides an example 
of a case where the establishment of requirements and safeguards governing a restoration action can allow 
actions to proceed immediately without the delays built into existing statutes such as NEPA. The law and 
EPA regulations spell out the procedures required for restoration and NEPA compliance is not required, 
due to the fact that the procedures developed were found to offer equivalent protections. The experience 
under CERCLA and the NCP, however, also presents a model of mistakes to be avoided, in that, although 
those requirements allow immediate implementation, the procedures and safeguards required for longer 
term actions have been widely criticized as excessive, causing excessive cost and delay. 

Under CERCLA, where there is an immediate need to proceed to prevent continuing harm, EPA may 
proceed immediately to implement a "removal action" without the extensive study and public 
participation required for a full "remedial action" that will achieve final cleanup.^" If a planning horizon 
of greater than six months is required, somewhat more participation and study is required, more 
equivalent to an EA. However, response other than investigation and monitoring is limited to actions 
costing less than $2,000,000 and lasting one year unless EPA finds that continuing timely action is 
otherwise necessary to protect health or the environment." In many cases, restoration will be achieved 
without further action. However, in other cases, a full Remedial Investigation and Feasibility Study 
("RI/FS"), involving study of alternatives often greater than entailed in an EIS, is required. In its 
promulgation of the first version of the NCP, EPA detennined that compliance with other federal laws, 
such as NEPA, was not required, but that these procedures created equivalent protections.^^ Other federal 
agencies pursuing cleanup pursuant to the NCP have consistently detennined that NEPA compliance is 
not required. 

The CERCLA determination offers a helpful precedent, not because that model should be replicated for 
species and ecosystem restoration, but because it presents a situation where an agency has determined that 
NEPA and other federal procedural requirements are inapplicable because the action at issue ( I ) involves 
restoration and (2) incorporates requirements that will prevent abuse. The CERCLA model for study and 
procedures, particularly for the RI/FS, however, is unduly costly and has been widely criticized for 
causing the undue costs and delay that undermine the statutory goal of achieving restoration quickly and 
effectively. The optimal restoration model would incorporate the CERCLA model of establishing 
safeguards to achieve the goals of permitting rapid and effective restoration and excluding these actions 
from procedural requirements such as NEPA, while not creating a whole new set of requirements for 
study and procedure that will create greater delay and greater cost. 



42U.S.C. §§9601-9675. 

40 C.F.R. pt. 300. 

40 C.F.R. §300.415. 

42 U.S.C. § 9604(c). 

United States Environmental Protection Agency. 1985. Preamble. The National Oil and Hazardous Substances 
Contingency Plan, 50 Fed. Red. 47912, 1985 WL 126730 (Nov. 20, 1985). In the 1986 Amendments to that 
CERCLA, Congress further spelled out where compliance with other laws was required, specifying that permits not 
be required, 42 U.S.C. § 9621(e) but requiring that applicable and appropriate substantive (but not procedural) 
requirements apply to final cleanups, id. § 9621(d). 



187 



Problems with Existing Mechanisms to Facilitate American Chestnut Restoration 

The existing categorical exclusions in Park Service and Forest Service policies should allow 
reintroduction and restoration to proceed using adaptive management techniques w ithout the need for an 
EIS or EA. However, there are a number of potential problems that could arise to inhibit that action. 
These problems could also inhibit other types of ecological restoration programs. 

The existing exclusions are far from uniform. Moreover, they are not uniformly applied in the field. This 
is likely due to the lack of any clear standards or uniform policy governing ecological restoration on 
federal lands. Without clear guidance, as under ESA, agency personnel will often fall back on the more 
traditional "wait and study" approach that will delay and may prevent restoration actions. 

There is also often a lack of trust that is exacerbated by a lack of clear, transparent standards. The public 
often mistrusts the government. Moreover, stakeholder groups mistrust one another. Without clear 
guidance indicated when and how a restoration action will take place, these groups may demand more 
process and study. These demands may further induce agency personnel to delay action and stud\ the 
problem further. The lack of trust may also generate litigation when the agency does proceed w ith a 
restoration action which raises a concern for some interest group. 

The lack of clear guidance and a clear policy favoring restoration action over the "no action" alternative 
will also affect the courts, in the absence of a clear articulated and clear policy, the courts will also fall 
back on the traditional model. In fact, the judicial reliance upon precedent may force such a result 
without clear guidance on an alternative model that has been adopted by the agencies. This lack of 
guidance will create inconsistent results, as have appeared in the application of the ESA reforms noted 
above. 

All of these threads are apparent in appeals from actions of the Forest Service. The lack of trust noted 
above has generated an explosion of litigation, including litigation against restoration programs. Between 
1997 and 2002, 3737 appeals from Forest Service action were filed before the United States Courts of 
Appeals, including 139 appeals challenging restoration programs and 97 challenging prescribed bum 
programs often necessary for ecological restoration and possibly necessary for American chestnut 
restoration (Malmsheimer et al. 2004). This explosion of litigation, in part, motivated Congress to pass 
the Healthy Forests Restoration Act^ to allow restoration activities employing logging in order to address 
the adverse impacts of years of fire suppression. 

It is entirely possible and even likely that the proposals of the American Chestnut Foundation to proceed 
with its American chestnut restoration on Park and Forest Serv ice lands using adaptive management 
techniques will be able to proceed without an EA or EIS under existing categorical exclusions and 
without controversy or litigation. It is a popular native species unlikely to arouse opposition. 
Nevertheless, there are some characteristics of the restoration that could generate controversy and. in turn, 
inhibit both reliance on the categorical exclusion and efncient, early restoration action. 

It is possible that some stakeholders could contend that the backcrossed American chestnut is a new 
species and possible, but less likely that those concerned about alien invasive species could object to the 
reintroduction. Such an objection bears some risk of bringing the reintroduction program outside of the 
Park Service's categorical exclusion, which applies to "noncontroversial" native species. However, the 
facts that oriental chestnuts and hybrids have been widely introduced in the past and that the current 
backcross is phenot>pically and largely genetically American chestnut make it unlikcl> that such an 
objection could prevail. 



-'' Pub. L. 108-148, 117 Stat. 1888 (Dec. 3, 2003), cocy(//£?c/a/ 16 U.S.C. §§ 6501-6591. 



88 



The foregoing possibilities raise greater concern for the restoration of the genetically engineered 
American chestnut, even though the genetic makeup of that species would be American chestnut, but for 
four genes. Genetically modified organisms have spawned significant controversy, particularly in 
Europe. Although the relatively insignificant modifications introduced by genetic engineering make it 
less likely that the blight-resistant American chestnut could be considered a "new" species, significant 
controversy could deter either agency's willingness to rely on categorical exclusions. Moreover, 
litigation is often a throw of the dice, particularly in the absence of clear written guidance. 

Preparation of sites for planting blight resistant American chestnuts and management activities may also 
be unnecessarily restricted by limitations on the categorical exclusions. Because American chestnut 
requires sun, site preparation may require canopy opening (i.e. logging), which could arouse opposition or 
litigation. The amount of canopy opening may be limited. Use of herbicides and fire management would 
assist establishment of American chestnut and suppression of competing species. Deer management will 
be required. Many planting sites might be restricted. For example, chestnuts frequently appeared on 
slopes. Some of these methods might be deemed to bring the restoration efforts outside of the Forest 
Service categorical exclusions. 

On balance, though, existing categorical exclusions should support a properly managed program, if 
constrained, restoration of the American chestnut. Nevertheless, there is a risk of some controversy. 
These risks and the often unnecessary limitations upon the existing categorical exclusions suggest that a 
more consistent and unambiguous restoration policy will be useful. These risks could generate even 
greater risks for other restoration efforts, which would also benefit from a firm interagency federal policy 
favoring action on restoration over inaction, embodied in regulation or clear policy cutting across agency 
jurisdictions and programs. 



An Alternative Model For Environmental Restoration Projects 

An alternative to the current model that would better advance the underlying intent of NEPA in 
addressing ecological restoration activities would treat the "restoration action" in the same manner as the 
"no action" alternative to treated in other cases, allowing that action to proceed while requiring a 
justification for failing to act. This alternative could be achieved by regulation or by policy, including a 
policy regarding categorical exclusions. It could also be achieved through appropriate legislation. 

While this approach is conceptually simple and consistent with the intent of NEPA, it becomes somewhat 
more difficult in application. The "no action" alternative is readily defined as "doing nothing" or 
"carrying on as usual." However, there is an infinite variety of positive actions that one could take. Thus, 
one must define what type of action constitutes a restoration action. This requires, initially, a definition of 
the goal of the action. It also requires definition of the quality of the action - - the action must be 
reasonably calculated to achieve that goal (Henry & Encash 2000-2001). 

A clear and somewhat limited definition of a restoration action is also necessary to prevent abuse. Many 
actions to advance interests other than restoration of the environment can be pretextually labeled as a 
"restoration" action. For example, many of the concerns regarding removal of accumulated fuel in forests 
arise from the fear that other forest harvests will be dressed up as fuel reduction programs where the real 
intent is to maximize profit by maximizing the amount of wood harvested. On the other hand, going too 
far to address these concerns can result in excessive restrictions which undermine the intent of expediting 
restoration, as has occurred under the CERCLA program. 



189 



Finally, some limitation in the types of restoration action that should proceed rather than waiting and 
studying alternatives is warranted in cases where the restoration may have an adverse impact on other 
important values. In these cases, a collision with other values may fa\or a more deliberative, if costK 
process. For example, where predator reintroduction would threaten human safet) or threaten major 
impacts on economic interests, more study can be necessary before proceeding. 

Many of the differences in categorical exclusions found in existing agency procedures may be explained 
as each individual agency's response to the foregoing concerns. However, the existing approaches are ad 
hoc and often inconsistent. We need more consistent, focused approach. Some guidelines for such an 
approach are suggested below. 

The presumption that a restoration action should be treated as a "no action" situation and proceed would 
be triggered upon a finding that the procedure has certain characteristics. The categorical exclusions 
established by the Fish and Wildlife Service for restoration projects and the Park Serv ice Guidance on 
restoration projects*'" include criteria that might be applied. The first such criterion should relate to the 
objective of the project. A restoration action should have the goal of restoring a native species or habitat 
that has existed in the site during historic times that has been removed or adversely affected b> human 
activities or the results of human activities. Restoration should not, however, be limited to cases where 
the hami has already occurred. It should also include actions to protect native species and habitats from 
harm that is reasonably expected. For example, removal of vegetation infected with sudden oak death (a 
disease that may also threaten chestnut restoration) should proceed immediately rather than waiting and 
studying the impact of removal. This should include efforts to allow systems to adapt to human induced 
changes that are inevitable, such as the changes that will likely be associated with human induced climate 
change. Because systems are in a state of change even absent human influence, the exception to the wait 
and study presumption should, initially be limited to restoration of ecosystems or species that were native 
in historic times. Whether restoration megafauna of the type that became extinct in the Pleistocene 
extinction, an issue that may face Park managers in the future (Flannery 2001, pp. 345-346), would be 
included should await development of further knowledge. 

A second criterion should address the manner in which the action will be carried out. To proceed without 
an EIS or EA, the restoration action must be carried out in a manner consistent w ith accepted practices, 
given the present state of knowledge. This does not mean that success must be assured. Knowledge of 
restoration is necessarily limited. For that reason, to qualify for the "categorical exclusion, the restoration 
action should employ adaptive management. Procedures for monitoring, reassessment and adjustment 
should be in place. In other words, procedural mechanisms must be in place to assure use of the best 
science available rather than a determination of the answers before action. 

Finally, limitations should be included to assure that restoration action will not create major adverse 
disruption of important or valued human or natural systems. The establishment of limitations presents the 
greatest challenge. Arguably, it is overprotective limitations that currently inhibit restoration actions and 
that have created m?ny of the inefficiencies in the CERCLA program as well as the ESA. Any limitation 
should have a defined threshold and not use vague terms that pro\ idc insufficient guidance to courts and 
agency person. For example, federal restoration actions should be subject to the same t> pes of controls 
that would apply to non-federal lands. Annual limitations on the size of areas in which all \egetation 
would be removed might constitute another type of limitation. The policy should be clear, however, that 
the presumption would favor restoration and that doubt should be resolved in favor of proceeding rather 
than vice versa. In some cases, time or cost thresholds may be appropriate, as is the case w ith CERCLA 
removal actions. 



''" NFS Management Policies, id. § 4.4.2.2. 



190 



Thus, a qualified restoration action would be one where ( 1 ) the project was designed to restore a system 
or species that had been removed or adversely affected by human disturbance or other disturbance related 
to human activity, (2) the restoration must be reasonably calculated to address or reduce effects of that 
disturbance (Henry & Lucash 2000-2001 ). and (3) the project is supported by a reasonable management 
plan that, if implemented, will not result in a serious threat to human safety, other environmental 
resources or property that cannot be adequately compensated with money damages, (4) procedures had 
been established for monitoring and adaptive management, and (5) the project did not any specifically 
established limitation or threshold established to conserve other important values. 

This approach is consistent with the recommendation of the NEPA Task Force Recommendations to the 
Council on Environmental Quality (CEQ 2003). That report recommended establishing an adaptive 
management work group to broaden use of that tool in NEPA implementation. It recommended 
broadening use of categorical exclusions, while incorporating monitoring and adaptive management to 
gather data regarding categorical exclusions. It also recommended better integrating NEPA into other 
programs including the Endangered Species Act consultation program. Adopting a consistent, 
interagency approach to restoration would be consistent with all these recommendations. It would also 
better effectuate the intent of NEPA that federal actions proceed in a manner that will protect and enhance 
the natural environment and encourage proactive, privately led programs such as the American chestnut 
restoration. 



Bibliography 

Adams, C. et al. 1998. Stream Corridor Restoration: Principles, Processes and Practices, USDA et ah 

Baltimore Gas & Electric Co. v. Natural Resources Defense Council. Inc., 462 U.S. 87. 103 S.Ct.2246 
(1983). 

Bean, M.J. 2003. The ESA - Second Generation Approaches to Species Conservation, Challenges to 
Making Second Generation Approaches Work. American Bar Association Section of Environment, 
Energy and Resources 1 1"^ Section Fall Meeting. American Bar Association. Chicago, IE, p. 883-892. 

Buniham, C.R. 1991. A Minnesota story: Restoration of the American chestnut. Reprinted in J. Am. 
Chestnut Found. 17(1): 16-27 (2003). 

Buttrick, P.L. 1915. Commercial uses of chestnut. Reprinted in J. Am. Chestnut Found. 13(l):21-28 
(1999). 

Center for Biological Diversity v. United States Fish and Wildlife Service, 202 F. Supp. 2d 594 (W. D. 
Tex. 2002). 

Council on Environmental Quality. 2003. The NEPA task force report to the Council on Environmental 
Quality: modernizing NEPA implementation, http://ceq.eh.doe.gov/ntf/report/finalreport.pdf 

Council on Environmental Quality Regulations, 40 C.F.R. Parts 1500-1517. 

Craddock, J.H. 2000-2001. The American Chestnut Foundation Symposium on Species Restoration. 
Journal of the American Chestnut Foundation 14(2):7-9. 

Endangered Species Act, 16 U.S.C. §§ 1531-1544. 



191 



Endangered Species Act Amendments of 1982, Publ. L. No. 97-304, 1982 U.S.C.C.A.N. (96 Stat.) 1411, 
1422, § 6; and ll.R. Report No. 97-567, at 17. 33-35. reprinted at 1982 U.S.C.C.A.N. 2807, 2817. 
283333-2835. 

Federal Insecticide, Fungicide and Rodenticide Act ("FIFRA"), 7 U.S.C. §§ 136-136y. — ^^ 

Federal Land Policy and Management Act, 43 U.S.C. §§ 1701-1785. ^_- 

Flannery, T. 2001. The eternal frontier: An ecological histor> of North America and its peoples. Grove 
Press, New York, New York. 

Frelich, L.E., and K.J. Puettman. 1999. Restoration ecology. Maintaining biodiversitv in forest ^^ 
ecosystems. Hunter, M.L. (ed.). Cambridge University Press. 

Gerber v. Norton, 294 F.3d 1 73 (D.C. Cir. 2002). 

Gjerstad, D. 2000-2001. The longleaf alliance: a regional restoration effort. J. Am. Chestnut Found. 

14(2):24-27. 

Harker. E.. G. Libby. K. Harker, S. Evans, and M. Evans. 1999. Landscape restoration handbook. 2nd 
edition. Audubon International Lewis Publishers, New York. 

Hebard, F.V. 2001. Meadowview notes 2000-2001. J. Am. Chestnut Found. 15(1):7-I7. 

Hebard, F.V. 2002. Meadowview notes 2001-2002. .1. Am. Chestnut Found. 16(1 ):7-18. 

Hebard, F.V. 2003. Meadowview notes 2002-2003. J. Am. Chestnut Found. 17(1):7-14. 

Henry, V.G. and C.F. Lucash. 2000-2001. Species restoration - lessons from Red Wolf reintroductions. J. 
Am. Chestnut Found. 14(2):35-39. 

Irwin, H. 2003. The road to American chestnut restoration. J. Am. Chestnut Found. 16(2):6-13. 

Klinger, C. 2000. Starting chestnuts in the forest. Northern Nut Grower's Association 9r' Annual Report, 
reprinted in J. Am. Chestnut Found. 15(2):40-45. 

Kriclier, J.C. and G. Morrison. 1988. Eastern forests. Houghton Mifflin Company, Boston. 

Leslie. M., G. Mcffe, D. Hardesty, and J. Adams. 1996. Conserving biodiversity on military lands - A 
handbook for natural resource managers. Department of Defense - Biodiversity Initiative, The Nature 
Conservancy. 

Lord, W. 1998-1999. William Lord's wildlife connection essays. Reprinted in J. Am. Chestnut Found. 
13(2): 12-16 (1999-2000). 

Malmshcimer. R.W., D. Keele, and D.W. Floyd. 2004. National forest litigation in the US courts of 
appeal. J. For. 102(2):2O-25. 

Morgan, J.J. and S.H. Schweitzer. 1999-2000. I lie importance of the American chestnut to the eastern 
wild turkey. Journal of the American Chestnut Foundation 13(2):22-29. 



192 



National Environmental Policy Act , 42 U.S.C. §§ 4321-4370f. 

National Forest Management Act, 16 U.S.C. §§ 1600-1614. 

National Oil and Hazardous Substances Pollution Contingency Plan, 40 C.F.R. pt. 300. 

National Park Service Organic Act, 16 U.SC. § 1. 

Oak, S.W. 2002. From the Bronx to Binningham: Impact of chestnut blight and management practices on 
forest health risks in the Southern Appalachian Mountains. J. Am. Chestnut Found. 16( 1 );32-41 . 

Paillet, F.L. 2002. Chestnut ecology-A personal perspective. J. Am. Chestnut Found. 15(2):20-31. 

Perry, J. and C. Ison. 2003. The impact of fire on chestnut in the central hardwood region. J. Am. 
Chestnut Found. 17(1):34-41. 

Rhoades, C.C. 2001. Pre-blight abundance of American chestnut in Kentucky. J. Am. Chestnut Found. 
l5(l):36-44. 

Ruhl, J.B. 2004. Past, present, and future trends in Endangered Species Act. Public Lands and Resources 
Law Review, accepted for publication. 

Russell, E.W.B. 1987. Pre-blight distribution ofCastanea dentata (Marsh.) Borkh. Bull. Torrey Bot. Club 
114(2):183-190. 

Taylor, M.E. 2002. Moving away from command and control: The evolution of incentives to conserve 
endangered species on private lands. Proceedings of the 3"^*^ Goddard Forum. The Pennsylvania State 
University. To be published in McKinstry, R.B. (ed.). Biodiversity: Addressing a global issue locally: A 
policy perspective on what biodiversity is, why we care, and the shape and importance of state, local and 
private policies and programs. Environmental Law Institute. Washington, D.C. 

Teich, G.M.R., J. Vaughn, and H.J. Cortner. 2004. National trends in the use of forest service 
administrative appeals. J. For. 102(2): 14- 19. 

United States Department of the Interior, Announcement of Final Policy for Candidate Conservation 
Agreements with Assurances, 64 Fed. Reg. 32, 726 (June 1 7, 1999). 

United States Department of the Interior, Announcement of Final Safe Harbor Policy, 64 Fed. Reg.32,71 7 
(June 17, 1999). 

United States Department of the Interior, Department of Interior Manual. National Environmental Policy 
Act Implementing Procedures for the Fish and Wildlife Services, 516 D.M. 6, Appendix 1, 62 Fed. Reg. 
2375 (January 16, 1997). 

United States Department of the Interior, Department of Interior Manual, National Environmental Policy 
Act Implementing Procedures for the National Park Service. 516 DM 6, App J, found at 
http://elips.doi.gov/elips/release/35 1 1 .htm . 

United States Department of the Interior, National Park Service. 2000. Management Policies 2001. 
NPSD1416. 



193 



United States Department of the Interior, National Park Service. 2005. Draft 2006 NPS Management 
Po\\c\es, found al http://parkplanninu.nps.gov/document.cfm?pr5oiectld=13746&dociimentlD=12825 . 

United States Environmental Protection Agency. 1985. Preamble to the Promulgation of the 1985 
National Oil and Hazardous Substances Contingency Plan, 50 Fed. Red. 47912, 1985 WL 126730 (Nov. 
20, 1985). 

United States Fish & Wildlife Service, Proposed Revisions to the Regulations Applicable to Permits 
Issued Under the Endangered Species Act, 68 Fed. Reg. 53327 (May 3, 2003). 

United States Forest Service, Forest Service Handbook, FSH 1909.15-92-1, Environmental Policy and 
Procedures ^ 3 1 .2(5), found at http://\v\\\v.fs.fed.us/emc/nepa/includes/epp.htm#c31 -^_ 

Wilderness Act. 16 U.S.C. §§ 1131-1 136. 

Wright, J., and G.L. Kirkland. 1999-2000. A possible role for the chestnut blight in the decline of the 
Allegheny woodrat. J. Am. Chestnut Found. l3(2):30-35. 



194 



Steiner, K. C. and Carlson, J. E, eds. 2006. Restoration of American Chestnut To Forest Lands - Proceedings of a 
Conference and Workshop. May 4-6, 2004, The North Carolina Arboretum. Natural Resources Report 
NPS/NCR/CUE/NRR - 2006/001, National Park Service. Washington, DC. 

FEASIBILITY OF LARGE-SCALE REINTRODUCTION OF CHESTNUT 

TO NATIONAL PARK SERVICE LANDS: SOME THOUGHTS 

. Scott E. Schlarbaum', Sunshine L. Brosi', and Sandra L. Anagnostakis' 

"^ Department of Forestry. Wildlife, and Fisheries, The University of Tennessee, 

274 Ellington. Knoxville, TN 37996 (tenntip@utk.edu) 
" Department of Plant Pathology and Ecology, The Connecticut Agricultural Experiment Station, 

123 Huntington Station, New Haven, CT 06504-1 106 



Abstract: In the early 1900s, chestnut blight transformed the eastern deciduous forest by eliminating 
American chestnut as a dominant overstory species. Currently, reintroduction of American chestnut 
appears possible with blight-resistant chestnut hybrids soon becoming available from various breeding 
programs. Overcoming chestnut blight through integrating resistance from Chinese and Japanese 
chestnut into the American chestnut genome represents a crucial first step in the process of restoration. 
Successful reintroduction, however, requires consideration of site, seed source, seedling quality, and 
silvicultural requirements. In addition, there are other challenges in eastern forests that await resistant 
seedlings. Considerations for reintroduction procedures and potential problems are discussed. 

Keywords: American chestnut / Chestnut blight / seedling establishment / reintroduction 

INTRODUCTION 

Reintroduction of blight-resistant chestnuts to eastern North American forests is one of the most 
anticipated events in natural sciences by the general public. Unique to chestnut is the formation of 
private nonprofit foundations whose mission is to develop planting stock resistant to the chestnut blight 
fungus [Cryphonectria parasitica (Murr.) Barr]. As predictions for the release of blight-resistant seeds 
and seedlings approach, attention should shift from blight resistance toward the ecological and 
silvicultural considerations that will also determine the success of reintroduction efforts. Detailed studies 
of forest ecosystems, e.g., Clements (1916), and North American forestry research (cf. Pinchot 1947) in 
general, were just occurring, as American chestnut [Castanea t/e/z/a/a (Marsh.) Borkh.] was being 
eradicated as an overstory species, so little is known about the establishment and growth of this species. 

The feasibility of large-scale reintroduction of blight-resistant chestnuts requires a greater understanding 
of many factors. Aspects of chestnut establishment success include site selection, material selection, and 
anticipation of establishment challenges. Given the dearth of information on artificial regeneration of 
American chestnut, reintroduction strategy on National Park Service (NPS) lands should encompass 
existing infonnation on establishment techniques on chestnut and closely relate genera, e.g., Qiiercm L., 
prior to or concurrent with the forthcoming release dates for hybrid material under the framework of 
existing natural resource management policy. This paper will consider some of the important factors that 
will affect reintroduction on NPS lands. 



REINTRODUCTION CONSIDERATIONS 

Site availability on National Park Service lands 

Restoration of chestnut on federal and State land bases will be affected by land management policy for 
each respective agency. The forest is resilient in tenns of response to disturbance, and the gaps left by 



195 



dead chestnuts have long been filled by other species. National Park Service lands are generally 
managed to mimic natural disturbance regimes, as opposed to USDA Forest Service lands that may be 
subjected to harxcsting. The first priority for reintroducing chestnut is to identify the appropriate sites for 
establishment. 



American chestnut is an intolerant species and requires a yet-undetermined amount of light to grow and 
successfully compete in eastern forests. Opening sites for planting blight-resistant chestnuts through 
harvesting is not a viable option for National Park Service lands. On larger NPS lands, suitable sites for 
planting chestnut can become available through the effects of natural disasters, fire, and pests on forested 
land. Tornados, straight-line winds, and hurricanes all can cause forest destruction b} blowing doun or 
shattering trees, thereby opening sites. Depending on the severit>', fire can remove the overstory and 
understory forest on significant acreages. Natural and exotic pests can also create openings and gaps in 
which chestnut could be successfully established. 

Seed Source Considerations 

Conservation programs, aimed at increasing disease resistance can reduce genetic variation u ithin the 
host species population. In the 1950s Forest Service breeding production programs for blister rust- 
resistant white pine in the western United States incorporated only 100 selected parent trees for 
reintroduction (Neuenschwander et al. 1999). The resulting seedlings, though disease-resistant, were 
limited in their 'fitness" or ability to adapt to varying environmental conditions (Buchert 1994). The 
natural geographic range and range of forest sites of American chestnut is large and therefore, individual 
breeding programs for different regions need to be implemented. 

The use of seed sources adapted to local conditions is an important factor in artificial regeneration success 
(cf Zobel and Talbert 1984). When information about genetic variation and adaptability of the species 
are unknown, locally-adapted seed sources have proven to be importance for seedling sur\ ival and mast 
production (Wakeley 1963). incorporating of locally-adapted genotypes into breeding programs is 
desirable, as the use of unproven genotypes can cause problems in: 

• Short-term and long-temi sur\ ival; 

• Poor adaptability to local environment can mean poor producti\ it> in tenns of growth and 
frequency and quantity of mast production; 

• Unproven genot>pes could affect overall productivity of the forest through pollination of 
naturally occurring trees of the same species; and 

• Planting of seedlings from unspecified seed sources could be challenged by private citizen, e.g., 
■■environmentalist."" groups 

As National Park Service lands are managed as bioreserves, use of genotypes from the local gene pool in 
a breeding program is particularly desirable. Practicalitx. however, is another issue, as resources for 
separate breeding programs for each National Park Service land base is not feasible in the foreseeable 
future. Therefore, the critical question is, how far can seedlings of one breeding program be mo\ed and 
still exhibit satisfactory survival and growth? In addition, there are significant en\ ironmental differences 
among chestnut sites within NPS lands with large acreages that can affect survival and growth. 

In general, there are few guidelines for seed transfer in hardwood species (cf Post et al. 2003) due to a 
relatively small amount of genetic testing done, in comparison to certain coniferous species. In tiie 
absence ofthe.se guidelines, identification of geographic areas with common en\ ironmental conditions, 
i.e., seed zones, can be useful to guide testing to determine seed transfer parameters. The wide range of 
physiographic and climatic conditions in Tennessee has lead to the creation of a hardwood seed-zone 



196 



system as a guideline for seed collection throughout the state (Post et al. 2003). This system was 
developed using a Geographic Infomiation System (GIS) and was based on elevation, Bailey's 
Ecoregions, 30 years of monthly precipitation, and 30 years of monthly minimum temperature data. 
Similar models can be created for large NPS land bases in combination smaller, proximal NFS land bases 
to guide seed transfer testing. 

It is possible that these models may be further refined by using a geographic information system (GIS) 
analysis that incorporates site information from historic and current chestnut sites. Butternut {Jugkms 
cinerea L.) is an eastern hardwood species that is being decimated by a disease caused by an exotic 
fungus (Sirococcus clavigig?ienti-jug/andacearum Nair, Kostichka, & Kuntz). Over 80 percent of the 
butternut population in southern States has been destroyed (USDA Forest Service). Surveys for surviving 
butternuts in the Great Smoky Mountains National Fark, Mammoth Cave National Fark, and St. Francis 
National Forest have been aided by predictive models generated by GIS analyses (van Manen et al. 2002; 
Thompson et al. 2004a. b). 

Seed Production 

Seed orchards will be necessary to generate the copious amount of chestnuts needed for reintroduction on 
NPS lands. Placement of seed orchards should be carefully coordinated with the NFS lands that they are 
intended to serve. Unpredictable expression (or lack of expression) of certain traits have been known to 
occur in progenies from seed orchards that were located remote to the intended areas of reforestation 
(Skroppa and Johnson 2000). Development of seed orchard management protocols will be another area 
that will demand attention. Management protocols for American chestnut orchards can be guided by the 
large volume of experience and research on several Castanea Mill, species (J.H. Craddock, Fers. Comm., 
May 1.2004). 

Development of pest management schedules specific for chestnut seed orchards will probably be needed, 
as North American pests are different from European and Asian pest species. Post et al. (2001 ) studied 
the effects of insecticide spraying in a Oiiercus rubra L. seed orchard and found that a number of seed 
pests attack acorns. A variety of other pest species were also found in this orchard (Schlarbaum et al. 
1998), which indicates that a significant amount of research will be needed to keep chestnut orchards 
healthy. 

In general, the NPS does not have the necessary expertise, suitable land, nor equipment for seed orchard 
development and management. Successful seed orchards require expertise in tree improvement, genetics, 
forest pathology, and forest entomology and well as a technical staff that is trained for working in highly 
managed conditions. Seed orchards require pesticide spraying and fertilization, which are not usually 
conducted on NFS lands. The NFS also does not generally have the type of equipment needed for seed 
orchard management. As seed orchard development becomes more eminent, the NFS should consider 
cooperating with USDA Forest Service Regional Genetic Resources Programs and State Divisions of 
Forestry (or equivalent), which have trained personnel in seed orchard management, equipment, and the 
land with appropriate variances for insecticide spraying, etc. 

Hardwood Seedling Quality and Establishment 

American chestnut has been planted since the mid 1 800s (cf. Emerson 1 846). By the 1 880s, experiments 
in establishing American chestnut on the Great Plains had been conducted for a number of years 
(Egleston 1 884). Detailed studies of growth after establishment are lacking from these early years, as 
forestry research, indeed forestry itself, was in an infantile state in North American during the latter stages 
of the 19"^ century (cf Finchot 1947). Because of the lack of information regarding chestnut seedling 
establishment, it is advisable to adopt guidelines from comparable hardwood species in order to develop 



197 



initial strategies to improve chestnut establishment. While survival rates and specific site and silvicultural 
requirements differ between oak species and chestnut, information on artificial regeneration of oak can 
provides useful information. 

Establishment of seedlings from heavy-seeded hardwood species through natural or artificial processes in 
eastern hardwood forests is becoming increasing difficult in recent years. Invasions of sites by exotic 
plant and vine pests and a dramatic increase in white-tailed deer {Odocoileus virginianus Zimmerman) 
herds augment difficulties with suppression by faster growing light-seeded species and hardwood sprouts. 
Artificial regeneration protocols are needed to combat existing and growing challenges for successful 
establishment of blight-resistant seedlings on NPS lands. 

Production of large, vigorous seedlings is a partial answer to these problems. Seedlings that are able to 
either maintain or grow sufficiently to keep the temiinal bud/shoot above competition and above the level 
of deer browse will have an increased chance for survival and establishment. Nurser\ protocols 
developed by the USDA Forest Service's Institute of Forest Genetics and implemented b\ the Georgia 
Forestry Commission at the Flint River Nurser> (Kormanik et al. 1993) have been shown to produce high 
quality 1-0 hardwood seedlings that are taller, thicker and have more robust root systems hardwood 
seedlings produced under standard nursery protocols. Studies with high-qualit> oak seedlings show a 
relative increase in establishment success (Kormanik et al. 2002). Chestnut establishment success will 
probably increase if seedling size is optimized. 

Chestnut seedlings have responded favorably to protocols developed by Kormanik et al. (1993) at the 
Flint River Nursery and to an additional season of growth at in the northern Connecticut State Nursery . 
Some American chestnut seedlings (1-0) grown at the Flint State Nursery reached over 1 .8 m in height, 
while some chestnut hybrid seedlings (2-0) were approaching 2 m tall at the end of a second growing 
season in Connecticut. There were genetic differences in nursery height, root collar diameter, and number 
of lateral roots among genetic families at both locations. 

Plantings of the Georgia-gown seedlings in Kentucky incun^ed severe mortality and heavy competition 
(Brosi 2001 ). Mortality was primarily due to Phytophthora cinnamomi Rands., an exotic root rot disease 
brought into the countr\ in the early 1800s {teste Clinton 1913) and chestnut blight, which had infected 
some seedlings in the nurser> . HerbivoPy by deer also impacted the unprotected trees. Despite these 
problems and heavy competition from yellow-poplar {Liriodeiidron tulipifera L.). the sur\ iving trees 
grew on average 40 cm a year. After two growing seasons, the seedlings had doubled their average initial 
height (average height: 106 cm in 2000. 203 cm in 2002, and 224 cm in 2003). Three growing seasons 
after outplanting some seedlings reach almost 4 m tall and were competing with yellow-poplar sprouts 
and seedlings. 

Plantings of the Connecticut-grown seedlings in Connecticut incurred little mortalit\ . Competition was 
restricted with the use of herbicides and hand-cutting, and the trees were protected from herbivory with 
plastic mesh tree shelters. The seedlings (from hand-pollinated crosses) were second-backcross Japanese 
(Caslanea crenatu Sieb. & Zucc.) hybrids crossed with two different American Chestnuts, and Hrst- 
backcross Chinese (Caslanea nio/lissima Blume) hybrids crossed w itli the same two Americans. Growth 
was best in three forest clearcuts. and was poor in an old field, .lapanese h\ brids averaged 258 and 225 
cm height after three seasons in clearcuts and Chinese hybrids averaged 216 and 227 cm. in the old field 
Japanese hybrids averaged 213 and 1 19 cm and Chinese hybrids averaged 155 and 188 cm after three 
seasons. 

There were differences in seedling growth across nitrogen content in tvvo planting locations indicating 
chestnut's ability to respond to soil differences in both Connecticut and Kentucky. Chestnut is often 
considered a species that can grow across a wide range of nutrient conditions given its historical range. 



198 



Delineating nutritional factors that influence chestnut growth will provide better information for site 
selection in chestnut restoration and will allow for appropriate fertilization of seedlings to augment 
growth. Further investigations are needed into different soil conditions and nutrient amendments to 
detemiine their interactions with initial seedling growth. 



ADDITIONAL CHALLENGES AFTER RESTORATION 

Production of blight-resistant, timber-type chestnuts is an important milestone in the restoration of this 
genus to eastern North American forests. Resistance to chestnut blight disease is, however, only the first 
step toward restoration. There will probably be other challenges to chestnut throughout its life cycle 
from both native and exotic organisms. Indigenous pests such as the two-lined chestnut borer could 
emerge as serious problems, depending on the density and vigor of the restored species. Chestnut blight 
disease is just one of the serious exotic problems that planted chestnuts will face. Below is a list of some 
of the most serious exotic pests that may affect the plantings. 

Phytophthora ciimamomi 

Historically, mortality of American chestnut and Allegheny and Ozark chinkapins (Casfanea piimila Mill, 
and C. ozarkensis Ashe., respectively) from root rot disease caused by Phytophthora cinnamoini is second 
only to chestnut blight disease. As mentioned above, the disease entered the country in Georgia during 
the early 1800s and rapidly spread. By 1878, American chestnuts in North Carolina River basins were 
noted to be dying (Hough 1878). American chestnuts and chinkapins were essentially eliminated from 
wet, poorly-drained soils and soils with heavy clay content by the turn of the 20"^ century (Crandall et al. 
1945). Site selection is often considered the most important factor in reducing losses due to Phytophthora 
(Agrios 1997: Campbell and Copeland 1954) and will be an important consideration in planting blight- 
resistant chestnut. 

Chestnut Gall Wasp (Dryocosmus kuriphUus Yasumatsu) 

This insect was accidentally imported into the United States on smuggled budwood of Japanese chestnut, 
(Payne et al. 1975). The pest lays eggs in vegetative and floral buds, and feeding by larvae forms galls. 
Branch dieback can occur, and severe infestations can cause mortality. The range of chestnut gall wasp is 
still expanding north and west in eastern North America. 

Exotic Ambrosia Beetle (Xylosaudrns crasshisuhis Mot, and Xylosandriis saxeseni Blandford) 

These introduced insects have been found to infest chestnut seedlings and grafts in field and forest 
plantings and can cause mortality (Oliver and Mannion 2001 ). Other ambrosia beetle species have been 
recently imported into the eastern United States and could potentially cause problems for juvenile and 
adult chestnuts (Campbell and Schlarbaum 2002). 
European Gypsy Moth ( Lyman tria dispar L.) 

European gypsy moth will feed on American chestnut, but it is not a preferred species. 

Sudden Oak Death (Phytophthora ramonim Werres et al.) 

Sudden Oak Death was first detected in California coastal forests in 1995 (Werres et al. 2001 ), and has 
since progressed into Oregon forests. The disease causes mortality in a number of hardwood species, 
including Quercus and Lithocarpiis, which are in the same family (Fagaceae) as Castanea. Phytophthora 
ramorum has a wide host range and occurs on rhododendrons and other species grown by the nursery 



199 



industry. USDA APHIS did not initiate interstate restrictions on movement until 2002. Currently, it has 
been discovered in 14 other States including a number of eastern states on nurser\ stock exported from 
California and Oregon (F.T. Campbell. Pers. Comm., June 3. 2004). Tests have confirmed that European 
chestnut is susceptible to Phytophthora ramorum (Defra. 2004). Although this pathogen has been brought 
to eastern states on nursery stock, confinnation of infestation in natural or urban trees has yet to be 
reported. In addition to these species, the number of exotic pests is likely to increase in the forthcoming 
years due to trade agreements and treaties that fail to adequately protect the United States from new - 
threats (Campbell and Schlarbaum 2002). 



CONCLUSIONS 

The above text has elucidated some of the factors that will contribute to or challenge the successful 
restoration of chestnut. Although it may be a number of years before significant numbers of blight- 
resistant seedlings are available, much information can be gathered by planting pure American chestnut or 
advanced generation hybrid chestnuts. Chestnut blight disease does not necessarily infect and kill young 
seedlings upon outplanting. Seedlings can maintain good health and growth for a number of years and 
thereby, contribute to the understanding of the silvics of the species. Plantings of advanced generation 
hybrid chestnuts on NPS lands can always be cut down after their study objectives are fulfilled. Concerns 
about pollen/seed contamination into existing American chestnut gene pools on NPS lands are negligible, 
as Asian germplasm is not competitive in eastern forests (Schlarbaum et al. 1994). Such plantings also 
can provide a better understanding of sites likely to have Phytophthora cinnamomi, which will be ke> to 
future successful chestnut plantings. 



ACKNOWLEDGEMENTS 

Special thanks to the many cooperators who have helped make this research possible including: Paul 
Kormanik (USDA Forest Service's Institute of Forest Genetics); Tom Tibbs (retired USDA Forest 
Service); The Georgia Forestry Commission; Chuck Rhoades; Paul Vincelli (University of Kentucky); 
Fred Hebard and Paul Sisco (The American Chestnut Foundation); Mark DePoy and Brice Leech Jr. 
(Mammoth Cave National Park); Tom Hall (Pennsylvania Bureau of Forestry ); John Peny (Berea 
College Forest); Jeff Lewis (Morehead Ranger District of the Daniel Boone National Forest); Pamela 
Sletten (The Connecticut Agricultural Experiment Station) and Tree Improvement Program personnel 
(The University of Tennessee, Knoxville). 



LITERATURE CITED 

Agrios, G.N. 1997. Plant Pathology, 4"' Edition. Academic Press. San Diego, CA. 1 11 p. 

Brosi, S.L. 2001 . American chestnut seedling establishment in the Knobs and Eastern Coalfields regions 
of Kentucky. M.Sc. Thesis. Universit> of Kentucky. Lexington. KY. 66 p. 

Buchert, (j.P. 1994. Genetics of white pine and implications for management and conservation. For. 
Chron. 70:427-434. 

Campbell. W.A., and O.L. Copeland. 1954. Littlcleaf disease of shortlcaf and lohloll\ pines. USDA Tech. 
Rep. 940. Washington, D.C. 4 1 p. 



200 



Campbell, F.T., and S.E. Schlarbaiim. 2002. Fading forests II. Trading away North America's natural 
heritage? Healing Stones Found. Publ. 128 p. 

Clements, F.E. 1916. Plant succession: an analysis of the development of vegetation. Carnegie Instit. 
Wash. Publ. 242. 512 p. 

Clinton, G.P. 1913. Chestnut bark disease. Conn. Agr. Exp. Sta. Rep. 1912:359-453. 

Crandall, R.S., G.F. Gravatt, and M.M. Ryan. 1945. Root disease oi Castanea species and some 
coniferous and broadleaf nursery stocks caused by Phytophthora cimuimomi. Phytopathology 35:162- 
180. 

Defra: Department for environment food and rural affairs. 2004. Plants known to be susceptible to P. 
ramorum. http://www.defra.gov.uk/planth/newsitems/suscept.pdf April 14, 2004, 3p. 

Egleston, N.H. 1884. Report on Forestry. Volume IV- 1884. Government Printing Office, Washington, 
D.C.421 p. 

Emerson, G.B. 1846. Report on the trees and shrubs growing naturally in the forests of Massachusetts. 
Button and Wentworth, State Printers, Boston, MA. 547 p. 

Hough, F.B. 1878. Report upon forestry. Government Printing Office, Washington, D.C. 650 p. 

Konnanik, P.P., S-J.S. Sung, and T.L. Kormanik. 1993. Toward a single nursery protocol for oak 
seedlings. P. 89-98 hi Proc. of 22" South For. Tree Improve. Conf , South. For. Tree Improve. Conf. 
Comm. (eds.). USDA Forest Service, Southern Region, Atlanta, GA. 

Kormanik, P.P., S-J.S. Sung, D. Kass, and S.J. Zarnoch. 2002. Effect of seedling size and first-order 
lateral roots on early development of northern red oak on a mesic site: eleventh-year results. P. 332-337 in 
Proc. 1 1"' Bienn. South. Silvic. Res. Conf, Outcalt, K.W. (ed.). USDA For. Serv. Res. Pap. NE-144. 

Neuenschwander, L.F., J.W. Byler, A.E. Harvey, G.I. McDonald, D.S. Ortiz, H.L. Osborne, G.C. Snyder, 
and A. Zack. 1999. White pine in the American west: a vanishing species — can we save it? USDA For. 
Serv. 

Oliver, J.B., and CM. Mannion. 2001 . Ambrosia beetles (Coleoptera: Scolytidae) species attacking 
chestnut and captured in ethanol-baited traps in middle Tennessee. Environ. Entomol. 30: 909-918. 

Payne, J. A., A.S. Menke, and P.M. Schroeder. 1975. Dryocosmiis kuriphihis Yasumatsu, (Hymenoptera: 
Cynipidae), an oriental chestnut gall wasp in North America. USDA Coop. Econ. Insect Rep. 25(49-52): 
903-905. 

Pinchot, G. 1947. Breaking new ground. Harcourt, Brace and Co.. New York. 522 p. 

Post, L.S., F.T. van Manen, S.E. Schlarbaum, R.A. Cecich, A.M. Saxton and J.F. Schneider. 2003. 
Development of hardwood seed zones for the Tennessee using a geographic information system. 
Southern J. Appl. For. 27: 1 72- 1 75. 

Post, L.S., S.E. Schlarbaum, L.R. Barber, D.F. Tolman and R.A. Cecich. 2001. Capture (bifenthrin) 
reduces Curculio weevil damage in northern red oak acorns. J. Entomol. Sci. 36: 222-225. 



201 



Schlarbaum, S.E., S. Anagnostakis, and M.C. Morton. 1994. Evaluation of experimental chestnut 
plantings in eastern North America. P. 52-56 m Proc. Intl. Chestnut Conf. JuK 10-14. 1992. Morgantovvn. 
WV. 

Schlarbaum, S.E., L.R. Barber, R.A. Cox, R.A. Cecich, J.F. Grant, P.P. Kormanik. T. LaFarge, P.L. ' 
Lambdin, S.A. Lay, L.S. Post, C.K. Proffitt. M.A. Remaley, J.W. Stringer, and T. Tibbs. 1998. Research 
and development activities in a northern red oak seedling seed orchard. P. 185-192 in Diversity and ,- 
adaptation in oak species. Proc. 2nd lUFRO Genetics o^ Que re us meeting. K. Steiner. ed. 

Skroppa. T., and O. Johnson. 2000. Patterns ot adaptive genetic variation in forest tree species: the 
reproductive environment as an evolutionary force in Picea abies. P. 49-58 in Forest Genetics and 
Sustainability. C. Matyas (ed.). ^=^ 

Thompson, L.B., S.E. Schlarbaum, and F.T. van Manen. 2004a. A habitat model to predict butternut 
occurrence and identify potential restoration sites in Mammoth Cave national park. Final report. Fhe 
University of Teimessee, Knoxville, TN. 

Thompson, L.B., S.E. Schlarbaum, and F.T. van Manen. 2004b. A habitat model to predict butternut 
occurrence and identify potential restoration sites in the St. Francis National Forest, in preparation. 

van Manen, F.T.. J.D. Clark, S.E. Schlarbaum, K. Johnson, and G. Taylor. 2002. A model to predict the 
occurrence of surviving butternut trees in the southern Appalachian region. Chapter 43, p. 491-497 in 
Predicting species occurrences: issues of scale and accuracy, J. M. Scott, P. J. Heglund, M. L. Morrison, 
J. B. Haufler, M. G. Raphael, W. A. Wall, and F. B. Samson, (eds.). Island Press. 

Wakeley, P.C. 1963. How far can seed be moved? P. 38-43 w Proc. 7"' South. For. Tree Improve. Conf., 
Comm. On South. For. Tree Improve, (eds.). USDA Forest Service South. For. Exp. Sta., New Orleans, 
LA. 

Werres, S., R. Marwitz, W.A. Man in^t Veld, A.W.A.M. De Cock, P.J.M. Bonants, M. De Weerdt. K. 
Themann. E. Ilievea. and R.P. Baayen. 2001. Phytophthora rawonmi sp. nov., a new pathogen on 
rhododendron and viburnum. Mycol. Res. 105:1 155-1 165. 

Zobel, B.J., and J.T. Talbert. 1984. Applied Forest Tree Improvement. John Wiley & Sons, Inc. 505 p. 



202 



Steiner, K. C. and Carlson, J. E, eds. 2006. Restoration of American Chestnut To Forest Lands - Proceedings of a 
Conference and Workshop. May 4-6, 2004, The North Carohna Arboretum. Natural Resources Report 
NPS/NCR/CUE/NRR - 2006/001, National Park Service. Washington, DC. 

^^ EFFECTS OF PAST LAND USE AND INITIAL TREATMENT 

— ON CASTANEA DENTA TA SEEDLINGS 

.Jennifer E. Hewitt', Albert J. Meier-, John H. Stames-, Priscilla K. Hamilton-, and Charles C. Rhoades^ 

'Division of Science and Resource Management, 
Mammoth Cave National Park, Mammoth Cave, KY 42259 USA (Jennifer_Hewitt@partner.nps.gov) 
-Department of Biology, Western Kentucky University, Bowling Green, KY 42101 USA 
^Rocky Mountain Research Station, Fort Collins, CO 80526 USA 



Abstract: Efforts to impart blight-resistance to the American chestnut, Caslanea dentata, have yielded 
strains with some ability to resist the disease. However, in addition to Chestnut blight, root rot caused by 
Phytophthora is a major impediment to the re-introduction of the chestnut The degree of presence of 
Phytophthora has been associated with soil moisture and loss of forestland to agriculture. At Mammoth 
Cave National Park, the effects of ectomycorrhizal fungi treatments and anti-fungal treatments were 
tested on chestnut seedlings. In addition, the effects of placing seedlings in disturbed (e.g., agricultural) 
or undisturbed areas were analyzed. No significant difference (p=0.09) in survivability was found 
between disturbed and undisturbed sites. In addition, seedlings in disturbed plots grew significantly more 
in height than undisturbed plots (p=0.01 ), though with no difference in diameter (p=0. 1 7). The use of the 
fungicide Ridomil gold EC"" in conjunction with greater soil preparation was found to increase 
survivability in disturbed plots (p<0.0001 ), though not in undisturbed plots (p=0.76). Ridomil did not 
have a significant effect on growth. The effect of mycorrhizal treatments on survivability was not 
significant when compared to control treatments (p=0.91). 



INTRODUCTION 

The American chestnut, Castanea c/entafa, was once an abundant canopy tree of Eastern hardwood 
forests. It was highly valued by both wildlife and humans for its nuts and quality timber. The chestnut 
was often 25 percent of the forest or more, and grew to be one of the largest trees in its ecosystem type at 
up to 100 ft tall and 5 to 10 ft in diameter (Ronderos 2000). Mammoth Cave National Park is on the 
western border of the chestnut's native range, although recent searches have found a number of young 
trees and root-sprouts throughout the Highland Rim region of Kentucky and Tennessee. During the 
summer of 2003 a researcher found over 1000 small American chestnut trees in the Big Woods, an old 
growth area within Mammoth Cave National Park (Mark Depoy, Mammoth Cave National Park (pers. 
com,). A review of surveyed deeds between 193 1 and 1937 within the area of the current park found that 
chestnut was noted on 27% of the deeds and that chestnuts comprised 5.1% of comer trees (Rhoades 
2002). 

The demise of this stately tree began when a deadly fungus was introduced into the United States near the 
turn of the century. The fatal blight-causing fungus, Cryphonectria parasitica, was discovered in New 
York City in 1904, and was probably introduced on ornamental chestnut trees imported from Asia. With 
New York as the epicenter, the fungus spread rapidly throughout the eastern United States, and by the 
1950s, nearly all of the American chestnut trees in the 9 million acres of their natural forest habitat had 
been killed (Ronderos 2000, Smith 2000). Only a few large chestnuts and some shrubby, blight-infested 
trees survived the initial epidemic. 1 he blight only top-kills chestnuts, leaving the root system intact so 
that many trees re-sprout after the main stem dies. Although there are many of these sprouts left within 
the original range, they usually do not survive long enough or remain healthy enough to begin producing 
seed. 



203 



After the near loss of this keystone species, the search for bh'ght-resistant trees began, and resistant parent 
trees were bred together in the hopes of generating even more blight-resistant offspring. Researchers and 
arborists also began back-crossing American chestnuts with completely resistant Asian species to create 
blight-resistant hybrids. Both hybrid and 100% American chestnut breeding programs have progressed 
significantly toward their goals, but there is not yet a genetically pure tree that is completely resistant to 
the chestnut blight. Many universities and natural areas programs have begun planting the more resistant 
trees that are available in an attempt to reintroduce Castanea dentata to eastern forests, but there are many 
difficulties associated with this process. 

One of these difficulties is caused by another harmful fungus, Phylophthora cmnamomi, which was noted 
in trees even prior to the introduction of blight (Rhoades, 2003). It is thought to be associated with levels 
of soil moisture and with the clearing of forest land for agricultural use. Rhoades et al. (2003) found that 
chestnut seedling mortality was highest in wet, compacted soil such as would be common in areas highK 
disturbed by agriculture. 

In some studies, ectomycorrhizal fungi treatment has been shown to protect against root rot. Anti-fungal 
treatments have also shown promise in preventing the disease (Marx and Davey 1969). 

In the current study at Mammoth Cave National Park (MACA). 2,000 1 -year-old 100% American 
chestnut seedlings from the American Chestnut Cooperators Foundation were planted in various locations 
within the Park. We tested these trees for their growth responses to application of additional mycorrhizal 
fungus, fungicide treatment, and control (no treatment). We also tested for the presence of Phylophthora 
fungus within these treatment groups and analyzed a potential correlation between Phylophthora and tree 
mortality. 

In this study we hypothesized that the trees would have higher survival rates and more growth in 
undisturbed soils, and with either fungicide or mycorrhizal treatment. We also hypothesized that tree 
mortality would be associated with the presence of Phylophthora. 



METHODS 

Sites and Treatments 

We selected 20 sites within Mammoth Cave National Park for these plantings (Table 1 ). Plots were 
placed on north- or northeast-facing slopes since evidence has shown this ma\ aid the trees in resisting 
chestnut blight. The spring and fall freezing and thawing associated w ith south- to west-facing slopes 
accelerates the bark-splitting that increases the chance of fungal infection. In addition, the chosen sites 
were well-drained with a sandstone substrate and acidic soil type. All plots were placed in areas with 
relatively open under^tories and at least some canopy opening. We placed ten plots on soils which have 
historically experienced disturbance of the mycorrhizal layer {i.e.. agriculture), and ten plots in areas with 
no known histor\ of soil disturbance (including the Big Woods area). 

One hundred saplings from five different parentages were planted 2 m apart in a grid pattern w ithin each 
of the 20 m x 20 m plots. Each tree was marked with a metal tag with a unique identification number that 
was associated with its plot, specific location within plot, and family designation. We measured the 
height in centimeters (cm) and root crown diameter in millimeters (mm) of each tree at planting. On the 
plot grid, each row of 10 trees began on the downhill (northeast) end of the plot and ran upward to the top 
(southwest) end of the plot. After planting, vented hollies containing cotton balls soaked in co>ote urine 
were hung at the four corners of each plot to deter deer from browsing on trees. 

204 



Table 1 . Plot number, soil type, planting date, and area of plantings. 



Plot 


Soil Type 


Date 
Planted 


Location Description 


1 


undisturbed 


3/21/2003 


Houchens Ferry 


2 


disturbed 


3/25/2003 


Houchens Ferry 


3 


disturbed 


3/24/2003 


Houchens Ferry 


4 


disturbed 


3/25/2003 


First Creek, East 


5 


disturbed 


3/26/2003 


First Creek, East 


6 


undisturbed 


3/27/2003 


First Creek, West 


7 


undisturbed 


3/26/2003 


First Creek, West 


8 


disturbed 


3/28/2003 


First Creek, West 


9 


disturbed 


4/1/2003 


First Creek, West 


10 


undisturbed 


4/2/2003 


First Creek, West 


11 


disturbed 


4/2/2003 


Blue Spring Hollow 


12 


undisturbed 


4/3/2003 


Blue Spring Hollow 


13 


undisturbed 


4/4/2003 


Cubby Cove 


14 


undisturbed 


4/4/2003 


Cubby Cove 


15 


undisturbed 


4/7/2003 


Big Woods 


16 


disturbed 


4/8/2003 


Big Woods 


17 


undisturbed 


4/8/2003 


Big Woods 


18 


disturbed 


4/8/2003 


Big Woods 


19 


disturbed 


4/10/2003 


Big Woods 


20 


undisturbed 


4/11/2003 


Blue Spring Hollow 



Within each plot, we applied three different treatment t\'pes. One-third of the trees were root-dipped in an 
ecto-mycorrhizal gel and wrapped in wet paper for transportation to the planting sites. This gel was 
prepared by mixing 53 grams of DieHard"" Ecto Root dip with 1 13 grams of Horta-Sorb Sm"" water- 
absorbent gel and 18.93 L of water in a 5 gallon bucket. 

The fungicide treatment group was treated by watering in approximately 325 ml of a solution of 1 .3 ml of 
Ridomil gold EC"" in 1 1 .63 L of water at planting. In order to apply the fungicide we prepared a hole, 
broke up the soil, and watered the solution into the hole before planting the trees. 

The control group was planted with no treatment, but roots of both control and Ridomil-treated trees were 
wrapped in wet paper to keep the roots from drying en route to the planting sites. For efficiency of 
planting, these two groups were planted using a dibble or planting bar method. 

In order to keep treatments from mixing, treatment groups were arranged by row within the plot, with 
treatment rows chosen randomly. For example, row I may be treated with Ridomil (R), row 2 with 
mycorrhizae (M), and rows 3 and 4 control (C), etc. One row, also randomly assigned, was mixed (3R, 



205 



3M, 4C) so that the number of trees with each treatment would remain consistent throughout all of the 
plots. 

At the end of the 2003 growing season we checked the trees for surv ival and re-measured the height and 
root crown diameter of each tree. 

To test for presence of Phytophthora and mortality associated with it. we collected 188 trees from the 
plots in March 2004. We collected 141 dead trees and 47 live specimens and used an ELISA test to 
indicate presence or absence of fungus in the genus Phytophthoru. The test kits were ordered from the 
Neogen company and the tests were performed with assistance from the Biotechnology Center at Western 
Kentucky University. 

Statistical Design and Analysis 

We used survival, height, and diameter of the trees to assess which treatment and plot t\pe had the 
greatest success. We found the change in height and diameter by simply subtracting the height and 
diameter measurements at planting from the 2003 season end measurements. 
We tested for correlation between height change and diameter change using a Pearson correlation matrix. 

A Fisher's exact two-way (similar to chi-square) test was used to test for differences in the survival of 
seedlings in disturbed versus undisturbed plots. 

A two-way Pearson's Chi-Square test was used to test for survival of seedlings by treatment t>pe. 
One-way ANOVA were used to determine whether differences existed between the different treatments 
and agricultural versus non-agricultural sites. 



RESULTS 

We found an overall survival rate of 56% for the planted seedlings. Survivorship of C dentata seedlings 
was variable from site to site (Figure 1), with visibly higher survival rates in Cubby Cove (sites 13 and 
14) and the historically chestnut-rich area of the Big Woods (sites 15-19). 

Proportion Survivors By Plot 



> 

3 
(0 

C 

o 
o 

Q. 
O 



^ 



t-,. 



i i I J 



-i 



i 



2 3 4 5 6 7 e 9 10 11 i: 13 M 15 16 i: 18 19 20 



Plot Number 



Figure 1 . Survival rates for individual plots by plot number. 



206 



Percent survival by site type was 58.4% for disturbed sites and 54.5% for undisturbed sites (Table 2). 
Contrary to our initial hypothesis, \vc found no significant difference (p=0.087) in survival between 
disturbed and undisturbed sites using Fisher's exact test. 



Table 2. Frequency of alive or dead trees at the disturbed and undisturbed sites. No 
significant difference in values, Fisher exact test (p=0.087). 





Disturbed 


Undisturbed 


Total 


Alive 


584 


545 


1129 


Dead 


416 


455 


871 


Total 


1000 


1000 


2000 



To further test the differences between site types we calculated the amount of tree growth over the season 
and compared this between disturbed and undisturbed sites. Height difference and diameter difference 
were highly correlated (r- = 0.695) according to the Pearson correlation test, so because of that high 
correlation and the probability of error in the diameter measurements only height difference was used in 
the ANOVAs for difference between site types and treatments. 

Using a one-way ANOVA for change in height, we compared disturbed to undisturbed sites and found 
that the disturbed sites had a significantly greater increase in plant height than undisturbed areas 
(F= 10.55, df=l, p<0.001 ), as seen in Figure 2. Dead trees were measured as well, so desiccation and 
decay led to some negative change in height for both groups. This result also contrasted sharply with our 
initial hypotheses. 



5 


1 i 


0) 




o 






c 


t 


>. 


<U T 




\ 


l_ 




Q} 


\ 


51= 


\ 


■D 


\ 


••-' 1 


- \ r _ 


x: ^ 


iL 


."=" c. 


V 


<D ^ 


X i 

-1 


r 1 1 



disturbed 



undisturbed 



Site Soil Type 



Figure 2. Height difference from planting to the end of the first growing season on 
disturbed (D) and undisturbed (UD) soils (p=0.001 ). 



207 



After comparing the site types to each other we separated them and tested for differences between 
treatments within a site type. Table 3 shows the survival rate of trees for each of the three treatments on 
undisturbed sites. We found no significant difference (p=0.933) in proportion of sur\ iving trees for any 
treatment group within the undisturbed plots, indicating that neither fungicide nor m\corrhizal fungus had 
a strong effect on the survival of these seedlings. 

Table 4 shows the same data for the disturbed plots. The only treatment associated \\ ith a significant 
difference in survival was the Ridomil treatment group (highlighted) with p<0.001 . There are several 
possible reasons for this result related to planting method and site characteristics, which are considered in 
the discussion. 



Table 3. Frequency of alive or dead trees at the undisturbed sites versus the three 
treatments. No significant difference was found using the Pearson Chi-Square Test 
(p=0.933). 





Control 


Mycorrhizal 


Ridomil 


Total 


Alive 


184 


179 


182 


545 


Dead 


157 


151 


147 


455 


Total 


341 


330 


329 


1000 



Table 4. Frequency of alive or dead trees at the disturbed sites versus treatment. A 
significant difference in the Ridomil treatment group (p<0.001) is observed using the 
Pearson Chi-Square Test. 





Control 


Mycorrhizal 


Ridomil 


Total 


Alive 


173 


166 


245 


584 


Dead 


165 


164 


87 


416 


Total 


338 


330 


332 


1000 



The last analysis we ran tested for the presence oi Phy tophi honi fungus. Using the 1 88 collected trees, 
we took root scraping", per the kit directions and found an overall infection rate of 25.5%. To examine 
whether or not the infections were actually causing tree mortalitv . we compared the infection rate among 
the dead trees (28.4%) to that of the live trees collected ( 1 7.0%). Interestingly there was no significant 
difference between the two infection rates (chi- =1 .36. df=l, p=0.24). indicating that Phylophihora is 
present within the plots but is not a major cause or tree death. We also isolated the results for dead and 
live trees in disturbed plots, where more infection and mortalitv' was expected, and found the same non- 
significant result (chi^ =1.82, df=l, p=0.177). 



208 



DISCUSSION 

There were several challenges involved in this experiment which may have had an effect on tree growth 
in our plots. Sites were chosen for relatively open understories, but there was still some degree of 
competition from other forest species, particularly the fast-growing red maples (Acer nibruni) and yellow- 
poplars {Liriodendnm tulipifera). Some of the planted seedlings also fell prey to light browsing by deer 
in a few plots and leaf damage from insects in most of the plots. One challenge that limits our ability to 
compare treatments is the planting method. In order to apply the Ridomil properly we had to loosen the 
soil and water in the chemical, while the other treatment groups were slot-planted to expedite the planting 
process. This differential treatment may explain a great deal concerning the apparent significance of the 
Ridomil treatment discussed below. 

We initially hypothesized that the trees would grow better in undisturbed soil sites and would survive at a 
higher rate with the anti-fungal treatment, but these hypotheses were not supported by the data. We found 
that seedlings grew more in height at the disturbed sites and that the Ridomil treatment had a significant 
positive effect on those sites as well. This initially led us to believe that Phytophlhora was more of an 
issue in disturbed plots and that the Ridomil was mitigating that problem. When we tested for fungal 
presence, however, we discovered that this was not the case, since there was no significant difference in 
infection rates between living and dead trees including those on disturbed soils. A more probable 
suggestion is that the significance of the Ridomil treatment is caused by the difference in planting 
methods. While loosening the soil may not make a noticeable difference in light, undisturbed soils, it will 
have a greater impact on trees if the soil in the plot is compacted by disturbance and has higher clay 
content. 

Although the growth was greater on disturbed soils, there was no significant difference in survival rate 
between site types, showing that the trees in our experiment still grew in the undisturbed areas, but a 
slower rate than at disturbed sites. In the Big Woods (plots 15-19). we found our highest survival rates, 
and this is most likely because of the site quality. Although all of the sites were chosen with the same 
criteria, the Big Woods area is known for the historic presence of large chestnut trees and currently holds 
over 1,000 natural seedlings and stump-sprouts, so more study is needed to find out exactly why 
Castanea dentata is so prevalent in this area and not in others. 

We will continue to collect data on these trees as they mature, and larger trends may be discovered in 
future years. From this initial set of data, we see that undisturbed sites are not necessarily better habitat 
for chestnut plantings and that a more involved planting method and an initial fungicide treatment will aid 
tree growth within the first year. We also note that although Phytophlhora fungus is present within our 
plots, it is does not appear to be a major cause of tree mortality within the sites we measured. We plan to 
continue this research by testing the soils and leaf-litter from each plot to determine more specifically 
what site characteristics lead to higher survival and greater growth among seedlings. 



ACKNOWLEDGEMENTS 

We would like to thank Mammoth Cave National Park staff and volunteers and the Great Onyx Job Corps 
crew for helping with initial measurements and planting, and helping with data collection at the end of the 
season. 



209 



LITERATURE CITED 

Marx, D.H., and C.B. Kavey. 1969. The influence of Ectotrophic mycorrhizal fungi on the resistance of pine 
roots to mycorrhizae to infections by Phylophlhora cinnumomi. Ph}topathology 59:559-565. 

Planting Protocols for American Chestnut Restoration Project. 2002. Mammoth Cave National Park science 
and resource management. Unpublished Internal Document. - ^ ^^; 

Rhoades, C.C. 2002. Historic abundance of American chestnut at Mammoth Cave National Park. 

Rhoades, C.C, et al. 2003. Effect of soil compaction and moisture on incidence of Phytophthora root rot on 
American chestnut (Castanea dentata) seedlings. For. Ecol. Manage. 184:47-54. ^=e._ 

Ronderos, Ana. 2000. Where giants once stood: The demise of the American chestnut and efforts to bring it 
back. J. For. 98(2): 10-11. 

Smith, David. 2000. American chestnut: Ill-fated monarch of the eastern hardwood forest. J. For. 98(2): 12- 15. 



210 



Steiner, K. C. and Carlson, J. E, eds. 2006. Restoration of American Chestnut To Forest Lands - Proceedings of a 
Conference and Wortcshop. May 4-6, 2004, The North Carolina Arboretum. Natural Resources Report 
NPS,^CR/CUE/NRR - 2006/001, National Park Service. Washington, DC. 

POTENTIAL EXTENT OF AMERICAN CHESTNUT RESTORATION WITHIN THE 

NATIONAL PARK SYSTEM 

s"^V William A. Leliis 

USGS, Leetown Science Center, Northern Appalachian Research Branch, 
1 76 Straight Run Road, Wellsboro, PA 1 690 1 USA (wlellis@usgs.gov) 



Abstract: A total of 128 National Park Service units and affiliated areas totaling 1 .9 million acres were 
identified as being located within or closely bordering historic American chestnut range. Fifty two parks 
had historical records of American chestnut; 35 of which were aware of living remnants within their 
boundaries. There is widespread interest among park managers in restoring American chestnut if a blight 
resistant tree was available that met park objectives, was ecologically sound, and was approved by NFS 
policy. The strength of the Park Service in contributing to a chestnut restoration program may lie more in 
numbers, distribution, variety, and miles of parks rather than in surface acreage alone. These factors 
make the National Park Service uniquely positioned to contribute towards a rangewide restoration effort 
through education, interpretation, research, evaluation, and demonstration of emerging blight control and 
resistance technologies. 

Keywords: National Park Service / Park Units / American chestnut / Restoration Potential 

INTRODUCTION 

The National Park Service (NPS) is a bureau of the U.S. Department of Interior with responsibility to 
preserve natural and cultural resources within congressionally designated areas for the enjoyment, 
education, and inspiration of current and future generations. There are approximately 388 units within the 
National Park system that collectively contain more than 83 million acres, in addition, the NPS 
cooperates with numerous federal, state, tribal, and local governments, private organizations, and 
businesses to extend benefits of resource conservation and outdoor recreation to many other public areas 
throughout the United States. 

In May 2004, the NPS sponsored a workshop in Asheville, l^C to review the ecological significance of 
the American chestnut {Caslanea deutata) to eastern forests, update the status of efforts to develop blight 
resistant trees, and explore opportunities and risks associated with NPS involvement in future chestnut 
restoration programs. The initial purpose of this paper was to determine the number and size of NPS 
units located within historic American chestnut range so as to estimate the maximum possible extent of 
NPS involvement in a chestnut restoration program. Subsequent to the workshop, additional infonnation 
was collected from park units on managers' knowledge of chestnut occurrence and their attitudes towards 
restoration. This information can be used to develop strategies and options that best utilize the unique 
resources of the NPS in a cooperative chestnut restoration program while meeting mandates to protect and 
preserve all natural and cultural resources within the National Park system. 



MATERIALS AND METHODS 

Identification of Park Units 

National Park units located within or closely bordering historic American chestnut range were identified 
by overlaying American chestnut coverage data with NPS unit boundaries or point locations using 



211 



ArcGIS ArcMap 9.1 (Environmental Systems Research Institute. Redlands, CA). Eastern U.S. state and 
county boundary layers were obtained from The National Atlas of the United States of America (http:// 
nationalatlas.gov) and historic American chestnut coverage was obtained from Little's Range and FIA 
Importance Value Distribution Maps (USDA Forest Ser\ ice. Northeastern Research Station, 
http://www.fs.fed.us/ne/ delavvare/4153/global/littlefia/species_table.html). ' 

A list of NPS units in the eastern U.S. was compiled from entries in NPS ParkNet (http://www.nps.gov), 
the NPS Owner's Manual (National Park Foundation 2002), and a National Park guidebook (Scott and 
Scott 2002). Parks were identified as individual units if they appeared on the list of "384 NPS Units" 
(File dbo_NPSDirector> _places.xls; NPS NR-GIS Metadata and Data Store: http://science.nature.nps. 
gov/nrdata/) and they were not part of another NPS administrative group. Parks within an administrative 
group were listed individually if all units in the group were included in the list of 384 parks, but were 
lumped with their administrative group if the group reported acreage from holdings not included in the 
384 park list. For example. Roosevelt-Vanderbilt National Historic Site parks (ROVA) were treated 
individually, whereas National Capital Central parks (NACC) were treated as a single administrati\e unit. 

Park boundary coverage and additional unit information was obtained from the NPS NR-GIS Metadata 
and Data Store (http://science.nature.nps.gov/nrdata/). Appalachian Trail coverage was obtained from the 
Appalachian Trail Conservancy (http://wvvw.appalachiantrail.org). All data layer projections were 
defined as NAD 1983 UTM Zone 17N and park units in chestnut range were extracted using Arc Toolbox. 
Park acreage was obtained either from FY2004 data reported for each unit on ParkNet (http://v\\sw.nps. 
gov/parks.html; see "Facts: Acreage" under each unit's web page) or from boundary calculations of shape 
files obtained from the NPS NR-GIS Metadata and Data Store (http://science. nature, nps.gov/nrdata/). 

Questionnaire 

A questionnaire requesting input on the following six questions was sent to each park unit identified as 
being within or closely bordering historic chestnut range (choices are in parentheses): 

1 ) Were American chestnuts ever present in your park? (Yes. No. Unknown) 

2) Arc there any living American chestnuts in your park today? (Yes. No. Unknown) 

3) Would your park be interested in restoring American chestnuts if a blight resistant form was 
available? (Yes, No, Unknown) 

4) If so, for what purpose? (Ecological Restoration = large scale reforestation; Demonstration = small 
plots for research purposes; Education = individual trees for historic reference or public education) 

5) If so, how many acres or how many trees would you anticipate at full restoration? 

6) If available, and within NPS policy guidelines, would the park utilize an\ or all of the following 
products? (Pure American chestnut selected for blight resistance; American x Chinese chestnut 

h\ bridized for resistance but retaining American form; Genetically Modified American chestnut with 
genes inserted for blight resistance from a different plant) 

Questionnaire recipients were identified through a search for managerial or natural resource staff through 
the NPS People and Places Director) (http://data2.itc.nps.go\/npsdirector>/). Affiliated areas and trails 
were not sent a questionnaire nor were acreages determined for those units. 



RESULTS 

A total of 128 National Park units and affiliated areas were identified as being vsithin or closely bordering 
historic American chestnut range (Figure I). These included 91 parks totaling 1.729.730 acres found 
within chestnut range. 21 parks totaling 130.686 acres bordering chestnut range, and 16 affiliated areas 



212 



and trails with undetermined acreage wholly or partially within chestnut range (Table 1 ). Represented 
within this group were 53 Historical Parks and Sites, 19 Military Parks and Battlefields. 12 National and 
Natural Parks, 1 1 Memorials and Monuments, 8 Rivers, 7 Heritage Areas, 6 Trails, 5 Recreation Areas, 3 
Parkways, 2 Preserves, and 2 Seashores. Park units were found within historic chestnut range in every 
state except Delaware, Indiana, and Florida. Excluding affiliated areas, parks ranged in size from 1 to 
521,752 acres (mean = 16,61 1 acres; median = 700 acres). The largest unit. Great Smoky Mountains 
National Park, accounted for almost one third of the total acreage found within historic American chestnut 
range. 

Questionnaires were sent to 104 of the 128 identified units, of which 81 parks (78%) responded; 63 parks 
within chestnut range and 18 parks bordering chestnut range (Table 1 ). Of the 63 parks located within 
chestnut range, 44 parks (70%) had historical records of American chestnuts on park land, of which 32 
parks (51%) were aware of remnant trees currently surviving on park grounds. Of the 1 8 parks bordering 
chestnut range, 8 parks (44%) had historical records of chestnut, of which 3 parks ( 1 7%) were aware of 
present day specimens. A total of 53 parks indicated interest in restoring American chestnut, while 1 1 
were not interested and 1 7 were unsure at this time. Of those parks expressing an interest, 3 1 would 
restore for ecological restoration of forests, 38 for demonstration and research purposes, and 34 for public 
education (parks could respond to more than one category). An additional 7 parks expressed an interest in 
using American chestnuts within historical or cultural landscape restoration, which was a category 
omitted from the questionnaire but should have been included. 

Forty parks identified a total of 79,441 acres for potential chestnut restoration and an additional possible 
need for 1,375 individual trees. These acres represent a total possible land area where American 
chestnuts could be incorporated into the forest or landscape, and no attempt was made to determine if this 
land was actually suitable for chestnuts or the tree density the land would support. Of the 63 parks that 
expressed an opinion on product type for potential restoration, all would accept a pure American selected 
for blight resistance. If a pure American was unavailable, 35 parks (56%) would consider using an 
American x Chinese hybrid and 35 parks (56%) would consider using an American chestnut with genes 
inserted for blight resistance from a different plant, assuming that use of the later two products were 
approved by NPS policy. Comments from individual parks on historical and present chestnut occurrence 
and attitudes towards restoration are listed in Table 1 . Some comments were paraphrased for brevity. 



DISCUSSION 

Recent advances in hybridization and genetic engineering have opened the possibility of controlling or 
managing the detrimental effects of chestnut blight in North America within the next several decades. 
Although these advances hold great promise to restore the American chestnut to its" former ecological 
role in eastern forests, the risks and benefits associated with using these technologies are poorly 
understood. Three paths are currently available to the National Park Service: 1 ) the NPS can chose not to 
be involved in chestnut restoration at this time and wait for other federal and private organizations to 
develop blight control technologies which the parks can implement after full testing; 2) the NPS can allow 
individual parks to adopt untested chestnut restoration technologies as they become available according to 
the level of risk each park is willing to assume, or; 3) the NPS can take an active, coordinated role in 
chestnut restoration by utilizing its unique resources to help other organizations and agencies develop and 
test new blight control technologies. Which direction the NPS chooses may well affect the ultimate 
success of returning this species to its native range during the next century. 

Response to this questionnaire indicates that a majority of park managers within historic chestnut range 
would like to restore chestnuts to their park landscape and would actively participate in a coordinated 
NPS restoration program. There was equal or greater interest in participating in a restoration program 



213 



through education and demonstration projects as there was for actual reforestation, and numerous parks 
expressed interest in incorporating American chestnuts into the cultural landscape. Although all 
managers would prefer using a pure American strain for restoration, there was no widespread objection to 
using an American x Chinese hybrid or a genetically modified tree if the product met park objecti\es. was 
ecologically sound, and was approved by NPS policy. 

This study identified over 100 parks and affiliated areas in historic chestnut range with total management 
area of approximately 1.9 million acres. Many of these parks have both historic and contemporary 
records of American chestnuts within their boundaries. Although this area represents less than 1% of the 
200 million acres that chestnuts once occupied in the eastern United States (see D.E. Davis, this 
proceedings for total historic acreage), these parks are widely distributed throughout all of historic 
chestnut range. Park lands may thus contain the full scope of habitat types once utilized by American 
chestnuts and remnant sprouts may contain the full array of remaining genetic variabilitv of the species. 
The strength of the Park Service in contributing towards an American chestnut restoration program may 
thus lie more in numbers, diversity, and geographic distribution of parks than in total surface acreage 
available for reforestation. This makes the Park Service uniquely positioned to cooperate w ith other 
public and private organizations to assist in chestnut restoration through public education, interpretation, 
demonstration, research, and long-term product evaluation. In addition, much of the acreage under NPS 
management exists in trails, parkways, and river corridors that traverse virtually all historic chestnut 
range, fhese units would be ideally suited to establish locally adapted source populations of blight 
resistant chestnuts for pollination and seed distribution into surrounding areas. 

The overall positive response to this questionnaire and enthusiasm for American chestnut restoration 
among park resource managers indicates the Park Service should move forward in a coordinated effort to 
develop a Service-wide restoration program. Possible steps include formalization of restoration policv 
including guidance on use of genetically engineered products in National Parks, creation of a committee 
or council to guide restoration efforts, identification of NPS research priorities and information needs, 
clarification of roles that specific parks could play in chestnut research and restoration, centralization of 
NEPA planning, cataloging of remnant chestnut resources on park lands for use in regionally adapted 
breeding programs, and establishment of formal relationships with universities, federal and state agencies, 
and private organizations interested in pursuing cooperative chestnut restoration programs. These steps 
would position the National Park Service to become a leader and showcase for American chestnut 
restoration, and could serve as a template for restoring other native trees facing similar ecological threats 
from invasive diseases and pests. 



ACKNOWLEDGEMENTS 

1 would like to thank Sara Fitzsimmons (TACF) and Jeff Cole (USGS) for assistance in creating the 
chestnut range map and identify ing parks, and Connie Johnson (USGS) for assistance in conducting the 
survey of park managers. 



LITERATURE CITED 

National Park Foundation. 2002. 2003 National Parks Pass Owner's Manual. National Park Foundation. 
Washington, DC. 65p. 

Scott, D.L.. and K.W. Scott. 2002. Guide to the National Park Areas: Eastern States, Seventh Edition. 
The Globe Pequot Press, Guilford, CT. 329 p. 



214 



Figure 1 . National Park Service units and affiliated areas located within historic range of American 
chestnut. Park numbers are identified in Table 1.^' 




Park areas within range 

Park areas outside range 

Historical range of American chestnut 



100 200 300 400 500 

Kilometers 



61 



A full color, PDF version of this map is available at: http://chestnut.cas. psu.edu/nps.hti-n#lellis 



215 



Table 1 . List of National Park units and affiliated areas within or immediately bordering historic American 
chestnut ranee ~. 







1 


1 


^ 


































































"O 




l-t 









0. 



































w 


'■^ 




















""" *> 


3" 

;7: 






y. 





> 

3 5 























V. 


■5 


5 


'J 


x; 


y. 


:= 


^ 




















'J 









^ 


:£ X 


c* 




















E 


^ 


'>^ 





E 




cL 








y. 










c o 
£ ^ 










•^ 


ra 


y 5 

'~3 *- 


i^ 




















r. 


■^ 





^ 


73 





•7: 


-3 


^ 




'J 










— ^ 

B ~ 

ll 

= op 
-3 = 







CO 

-V 

I* 




V-. 

c 

y. 

y, 
'_» 


1> 

C 


^ V 

3 ■= 
|| 

-a = 

_<: 5" 
c = 




Clj 

'J 

t3 


y. 
















'J 

r. 


y 

y. 
'J 

'-> 
y 


3 

>". 

■y. 

y 

C. 
y 


•J 
> 

•J 

■y 


g 
> 

y. 


5J 

■J 

c 
'J 

y. 

-3 
'J 




•6 

% 

y 


■y 

y 

y 

> 


re 

'd 

?3 
■y 
y 

y 

■y 
~ 

'-» 
•_* 

"c- 


> 

< 




y 

z 

y 

> 

-3 
< 


■y^j 








r3 ■*" 


o , 


;< 
"« 







Z 


2^ 


y. ^ 



























y 


1 


'_* 




•-> 

> 


y 


>-^. 


y 
.1^ 


"y 




CO 

3 


'0 










1 ^' 


^ 







S 




-3 


ra 3 




















— 





U 




c. 

'J 


y. 


■y 


=ij 


"3 






^j 


£ 

c 
U 








^ 2 


^ 5 

o u 








1 




•r. 


-y' 

> 
















^ 




> 
•J 


a. 


y 

■0 


■y 


1> 






y 




1 


u 
































































ul 










a; 












LlJ 












































e- 










r>' 












'K 




















— 










U-1 












^- 










^ 
































-^ 










d 












— ^ 
























C^ 








































^ 




































































"~~ 


































































C' ■ 
































































■o <U 


































































•=; -^ t 


































































§ i3 2 










> 












> 




















> 










> 












^ 


^-1 


































































>. r-. 
































































c c 












































































> 












> 




















^ 










> 












- 


3 1- 


































































u a- 


































































































































"ra ^. 






















u 










































'J 


1 8 

O iT. 










































'r 










D 












2 

5 










r- 












"/"i 










































»/', 


a. 






















~ 












































^_ 




































































































































































































rt '^^ 


































































o c 












































































> 












> 




















^ 










>■ 














































































■^ a. 


































































X 


































































































































o ^ 










cL 
























































V Q- 










— c 












• 




















m 










• 












' 


t- h' 










<■ "^ 












< 




















<• 










< 












<; 


iO 




























































^ 


r- 


■^ " 




-*• 






oo 






_„ 






__ 








































f^ 









a: 






'I- 






-T 













ri 






r- 


"/", 
•/", 




»/". 












00 










lyj 








7 






























£ 




-_ 
































^- 


























c 




__ 


V^ 


_ 
































>. 






■■J . 












D. 






z 


'_> 


J- 


Z 

-0 


z 


z 















y: 














'~f 




— 






2 






Q. 






z 






< 


z 


'J 

■r. 


>-. 


5 

z 


Z 










Z 










Z 


< 

2: 


« 




/; 






C 






z 






§ 






^ 


c jr 


'J 


S 


E 


— 


^ 












^ 










■3 


z 


z 

0. 




^ 






•a 






"0 






2 

1> 






/^ 


:|2 


--> 


2 


<- 

-3 


T3 


= 


.' 










--* 










"y 






? 












=^ 






5 






SJj 


c 




:^ 


-i: 





-3 


:5 













a 










U 






y 
















2 






^ 


c 


'J 


To 


*-» 


Clj 


^ 










TO 










-J_ 


"-'' 






^ 






ir. 




















■-r 


"^ 





■^ 
















ir 
















































"o 


c^ 


































Z' 


























•V* 




u 


-^ 


"_» 


~" 






























^ 


































z 




































— 






< 






< 






< 


< 


< 


< 


< 


<; 














> 










>- 


> — ^ 


V? 




^ 






Z 






^ 






^ 






^ 


^ 


^ 


2 


^ 


5^ 


s 













Z 










Z 


z z 


3t 


































































§■ 










1"^, 






^ 






^^ 






— 


r J 


-r 


'r 


^ 


r 1 


r^. 


'/^ 










^ 










r- 


X 




































'~l 


r J 


r J 










r^j 










rg 


(^^ 


s 














1 





















































216 









i> 










to 








CO 


^ 





C 


SI 


-u 
^ 












r 


^ 


1> 






















> 










3 






nil 


3 






a> 


':^ 












n> 


a> 






















z 

c 


5j 


> 

•J 

■•; 

■a 


o 


u 

'-^ '-) 
o c 

11 

£. °" 

^z 
H = 

C ^5 
ir^ ■ — 

:e -^ 

O '^ 
E C 

5 t: 




Z 

'y; 

>-. 

a 

I 

> 

a 
Q 


U 
,o 

3 
D. 


It 

'^ 


3 
E 

C3 

oi 



(/I 


CO 

Z 

'y 

I 
> 

Q 
2 


Xi 

3 

« 

p 


■0 

J 

3 

« 
01) 


3 

> 

3 

« 
"5 


73 


p 

V) 

ta 

Xi 

■a 
3 




n 

E 

'_» 

T3 

CO 

SI 

■o 
3 


■0 



3 
v^ 
CO 

->. 

CO 


■0 
3 

CO 

_E 
'J 





Oi) 

CO 

i 

3 


U 

x: 
xs 
-a 




'y 

0. 


"_< 

X5 

c 

v^ 
-■J 

cy 
It 
> 


c 

CO 
CO 

3 

V; 


D. 
>^ 

Q 

E 

a. 
3 

V. 


CJ 
CO 

-d 
u. OJ 

OJ 
OJ T3 

S g 


3 
OJ 

1) 

CO 
















E 
z 

T3 


Z 

s: 
'J 

r 

7Z 


C 
c 

X; 
73 


O 

u 
£ 


3 


z 

Q 
"o 


N 

-5 

0. 


3 
C 


-o 

^ § 

x: — 

1 5 
3 3 

a .0 


z 

1> 



U 

Z 
Q 




> 


3 
3 

SI 
•J 

3 

C3 


-a 

■u 

P 
u 


,0 
■0 

•a 
3 

■0 




-3 
3 
CO 

Q. 

Cl 

ra 
■a 
jj 


01) 

T3 

■U 
3 
CO 

— 


S 

ry 

V) 

CO 

x: 
U 


u 

Oi) 
>^ 

x: 

t— 

ry 



Q. 

3 
3 

CO 

C 
1 
E 


aj 

a: 
3' 


OJ 

CO 

C 

u. 

V5 

OJ 

CJ 

CO 


E "v? 


OJ 
CJ 

3 

VI 
OJ 

Qi 

■0 

CO 

c 



■3 








OX) 








O 


*-• 


n 


sz 




3 

c 

o 

o 

OJj 

fc 

< 
i 

c 


1» 


D. h 




li 




' 


ry, 

x: 


Oi) 
VI 


.^^ 


■0 



D. 

E 

i 

VI 

3 
CJ 

_o 

CO 






XI 




3 


"" 


'J 


^o 















i3 

c 

E 

E 
o 




2' 

z 

o 

■g 
'J 


Z 

o _ra 
11 


o 

(X 

z 

C 

i-U 
c/^ 

_3 


c 

■yi 

c 
c: 

*iZ 

o 

c 

OX) 

^> 
■a 


c a. 

a. 

-C 3 
OIJ v^ 

-5 -o 

C OiJ 

VI Xi 

!l 

o 
— -a 

4-* — 


h 

B « 

> a 

= 2 

- 2 
5 °- 

3 ^ 

is 

3 V, 




3 


3 

3 
P 
? 


u -O 

♦- 3 

^ s 

-o 

G ~ 

q: 

■0 * 
3 -^ 
.2 i 

« 


Q 



B 
c 

p 

■a 
(5 


1 « 

= i 

'J Oi 

£ £ 

E 3 
'J 


3 

a. 

vi 
■p 



OJ 



00 

■5 



'J 


1> 
'J 
3 

> 

3 

P 

3 
'J 




to 
c 



£ 

x: 
h- 

vi 

JJ 

CJ 

g. 

CO 
£ 

0. 


r;, 

t 
{ 


■0 
c 

L2 

V5 

p 

■0 



Q 



OJ 

3 

OJ 

■a 
> 


"§* 

'J 

c 


u 

a> 


a 

V. 


v: 
C 

E 
3 

VI 

E 

,0 

i> 

c 
v' 





Vi 
Q 

OD 

3 
'-> 

CJ 

3 
V5 


U 

C 
•-* 


"5 
« 
E 
■y. 



a. 
0. 

CO 
V5 

lj£ 

ro 

0. 

OJ 
3 


-— -d 
■y Ji 

1 i 

cB 

ii 

■^ i> 
M 1 


OJ 

n. 

OJ 

'.« 

OJ 

G 

"cj 

3 

1 
'-) 

2 









p 

Vi 

V) 
CO 

C 

CJ 

3 





o 
















UJ 




UJ 




tu 










































o. 

3 












UJ 




C 




Q 




^' 






















_l 
















Qi 




O. 




































































































































•^ -^ te 


































































o 59 2 












>- 




> 




>■ 




> 










z 












>- 






Z 










>- 




^'^l 


































































«^ 


































































c H 


































































ii 












Z 




^ 




-^ 




> 










z 












>■ 






Z 










> 




































































uc^ 














































































<n 






















































«^ -^ 












QJ 




































































































CO 

+ 






















Is 












O 




:^ 




ID 




D 




































CO 






O C/5 
0, 












'A. 


































— 
















f^l 




>, 


































































1 % 


































































" F 


































































'o S 












Z 




> 




>■ 




> 










D 












> 






2 










'^ 




.22 £ 


































































Si g. 












q 




~ 




I 




I 
























< 
















\^ 




Hti- 












< 




< 




< 




< 




































< 




if (5 

a. < 




2 


oo 


oe 











^C 


i^ 


- 


r 1 


r~^ 


- 




^ 










? 








Q 

c i/5 




Z 










;/5 






















c 
Z 








i/0 




:^ 




v. 






z 




Z 














Z 


















S 




OJ 




£: 


z 


1 

z 

S3 
a. 


Urn 

c 

> 




.r. 

i 
2 






c 

Xi 

o 

B 




'J 

O 

o 

U-1 




— I 
1 "^ 

^ 




'J 
3 
> 










i 
> 

c 

i 












Z 

CO 
OX) 



s 

CO 
C/5 






Z 

X 

c 

c 



u. 




5 


V5 -^ 

8 ^ 

JJ 3 
5.' 


Z 

•JO 

-5 

UJ 


Z 

3 

V) 

2 


c 
< 

OD 

■a 


i> 




































































y^ 




^ 




K^ "' 




,^ 




-^ 




^ 










^ 












K.^ 






^ 




,^ 


^ 


-. 


— . 


<r 


on 


z 




z 




z z 




z 




z 




Z 










Z 












Z 






Z 




z 


z 




z 


D. 


a. 


0^ 




_ 










rj 




f^ 




-1- 










■r, 












^ 






r- 




DC 


_ 


_ 


r^ 


^ 




ri 


















r^i 














r-^, 












i~^i 










^''i 


-t 


t 


-T 


T 


S 



































































217 



c 
o 
U 



— ^ o 

•— *" C 

i i = 
3 si 

5 2-" '5 



"3 


rj 






















x: 


^_ 


c 


<— 


















E 

1 » 


-J 

v. 




< 

It 


-a 




?: 


rt" 




a> 


T^ 


'-) 












c 


r3 


T3 




K 


1) 


i 


c 




-^ 




i> 


T3 




__ 



2 -3 s: s 



- :S 



U 





-a 




■a 

c 






!g 






t 








3 


c/o 


c 




O 





■^ -X '~> 



— g. -J 



3 u 






Vl CJ c 

_^ "^ ■;= 

^ — ~ 



.3 ;= 



1> 


r 


1* 








■n 


-a 


= 


-u 




■o 




C 


'!> 


1> 














'. > 


■o 


.~ 


rz 








o 




^ 


^ 




^ 



C3 l> C 



o 



•-» 3 ZJ 



^ 


g- 


o 
c 






















,„J 








— o 

— o 



■ — X Q 

w « n. 
"^ — .^ 

^ r^ ^ 



a 



^ x^ o 






-J '-> y. s ^ 5 — = o 

is ^ - a '■^' ^ '^ — 



<-i -o 



-3 2i -^ J^ _ 
= i sis 



a -J 



5 ;-; 2 o 



^ y 









- 5 < 



"2 € 






Cl 
3 ZJ 



c o 



9 -^ n - 



■;: ^r, = ra C 

2 7 fe - ? 



5^ g 



■a 

f3 "^ 






3J -^ 



-3 =!.« = -:=- 



^ <-■ i! E = 

-C — r^ 5 .= Q_ 
— w ^ =" . 






5J ^ :^ w — 



2 a 






C3 ^ "O -^ 












T, c 









5^ H 



1> U-, 

C3 u. -' 



« ,0 i2 h -= 



_ o. 
o 



S 3 

C [/5 



3 -^ ~~ ^ >". *-* 






— t c : ~ — 

.— O 1> — ''. 

^ 3 c ^i2 -5 

■5 = T3 -5 ^ 

C3 -y. c -^ -7^ 

'J t- S U "^ 

^ '-£•=; d 



5 g 

5 "O 



3 i^ 5 

■■2 ^ a:: 



^ := 



-3 


s~ 


v^ 


— 


-j; 


. 






'-» 

















u 


t^ 








< 


? 


^ 


^r. 


^ 


^— ' 


\r. 




;_; 












ra 






^ 



-3^-5 



? > ^ ^ 

y -3 =' 5 

S - 3 E 

>^. ^ _g = 

,^ "i c! ^ 



^ 'y- — w 



i i! t- — 

■y. 



— '7. > i i' 

;ij 77 < w c 

S c • S '^ _ 

■5 3 >> ?; ^ £ 



2 < B I i 



J-- ., '-^ zi. J-- 



- V. c - u 

_ = • c — ;2 
g = = " s 



£- :-! "2 






i . := 5 ^ 



.^ j5 ^ ^ t; 

^ < ^ g; j: 
=1 =P g -J £ I 



= ■= s ^ 






o ^ 



.- OC „ ir. ^ 

•3 T' S a £ 

T3 J. ^ C^ - 



5 c g -5 = 

o - ra = -_, 

T3 ^ - a t; 

? = - ~ s 

&•? g = ^. 






Q 

of 



Q 

of 



O ra *- 



a: 



>- 



u 0- 



?> 



C3 



U 



I 8 

O iT 

a. 



o o 



2 S3 






o- 



E 
a 

2 

(5 
c 



(3 



X 



1^ 



z 



q 
< 



00 



z 

Q. 

a 

u 
u 

Rl 



z 



o 
< 



a: 

a: 

on 

CO 

I 
u 

D 
b. 
u 
o. 

Q. 






a. 

z 

BO 



U 

O 



< 



218 



c 



o 



2 § 



O -3 



02 

C3 



^ N 



o — 

g o 

^ "7 

Si rA 



ii o 



9- |) M. 



N 



o 



a. 






U Z 



Di 2 



01} _y 



C c 






oij -if 

lU o 



CO 



c __ ^ 



- ^.i: y T3 15 



oo 

00 S 

c Z 



> 



y: ^ ,-^ 

. y c cf 

E u S 

— -a ■^ -J 



2 Z 



s 




ou 


a- 


c 


■o 


g 


u 




'« 


OJ 








'y; 




> 


Xi 


Q 


-a 


^ 


;/-■ 














P3 


— . 






■^ 


T3 

^ 


eg 


p.. 


> 


O 


-1 


■a 




c 






,o 


5 


E 






o 

z 


a 
Z 


"to 




QJ 


c 


a 


1» 


<J 


v 




r3 






"7 


.■s 


n. 
E 


u 



u 



•^ 



o 

u 



'J 

orj 






-3 ci: 



U CI. 



^^« 



z -a 

p o 






ojj 2 ■? = 



o 



'^ C ^ S 



E - 

of) — 

= r; 



o 



IJ 5 «- - 

•a « ^ 

g o ^ 2 

JJ o :' u C! n 

,~ -^ "^ « = ^ 

•^ -= a. -= ;:X a 



m .y 



-g § « 
5 13 - 



3 O 

> ■ ■ 



^. f- 



■B i 



S OX) 



2 " (u *- -^ 



2 c 

'^ 2 

Oi S 

c o 



, o 



a exj 

c 



II 



^ "3 



^ K 



,3 — 



= -u < 



i! — 



3 


n 


IJ 




^n 




c 


C 


c 
2 


"5 


1> 




o 


? 



U .S 1- 

-= E 

^ — 

(A £ 

u o. 

D. I— 

c/i O 



00 ■£ 

o .i£ 



P ^ — 



•S _ 



CO 



•u x: ■=: a 



o "- 
■i x: 



W ? l- 



c 



.5 
5 -u 

c ^ 

.2 o 



S -o ^-^ 



£ § 



_ -a 



' 3 
■a S ts 



; — _-. ^ 'yi 
I— < Oi) <u 



■ll E ^ 



_^ 



2 Qi 



la 



00 u 



o S 

1= C 






11-;-; TO 

t> C 'J 

« s y 

'T' .i: c/5 

^ X <J 



csi 



c 5 a 



3 
P3 x: ?3 



3 c 






s - 



_ ->-^ u 



M s 



— t- c 

E n 

2 £ i 

ojj< O 



u — 

E ^ 



.y ^ .2 

c -o 

U 1> KJ 

a _^ = 

t~- '^ Ob 



^ o - 

O „ 2 

.</> c c 

jr U O 

2 E O 

O c — 

"J fc 

1- Q Vj 

5 " 2 

< = 

a: o 

'Vi c 

z£ 



a 



Oil 

c ■ 



o£) ra 



01 



3 'O 



D. 
C/5 



u 



c 
o 
U 



Z 

O 



C 00 



U — c/i -= 



o -^ 
x: 00 



3 2 

o o. 

in ^ 



-O T3 

O u 

g s 



3 ii 

o c 



s.^ 



^ t 



° ^ - 



X V) 



3 5 



£:■ -a ,i; 

!^ ■= so 

fe o 2 

> 5 2 

^ 73 :2 

3 _«: 

Sis 



S -a i! i: 






'f Q. 



t- G 



VI 3 

!3 5 



- c ^ -5 

'-> t 00 — 

g ^ £ « ^ 

w «-t *T^ ■!> f^ 

^ w y n 



o 

3 



Q 






3 -E o 

o « -s 

^"1 



ID 






w 



u 



Is 

o vo 

ft. 



o 



o c 

•- «J 

O o 












a 



< 



< 



*5 t- 



o 



Z 



2 

Q. 



t/5 


ra 


I 


Qi 


z 


K 


^ 


00 


1> 


re i/D 


5 


K 3: 




a- z: 



< 

c 



m 
z 



u 
u 
u 

z 

■tt 

o 



< 

a. 



t/5 
X 

z 

o. 



z 

■a 



Z 



> 

00 



< 






219 





N 

re 






v. 




r. 

r. 
*-* 

2i 




■y. 


'y. . — r 

'•J ^ 

-^ s 

^ 1 


y. 




y. 


5 
'J 


"2. 


y 






'J 

u 
J3 


z 


'•A 
H 

r. 

-3 




•J 


y 

y 

— 


y 




y 

y 
y 


.5 


9 

y 


would be involved 
changin<; the 




y 
-y 


re 
'-> 

y. 

re 

y 





y 

> 


y 
y 


1 rcscarcl; iT 
iicnl Specialist) 




•J 

z 

c 
'-> 

■o 


--» 

::• 

o 
'J 

'J 

c 
re 

E 


I| 

c: 'J 

11 


c 


re 
re 

1l' 


z 

'J 

c 

z _ 


-T. 

b 
a- 

•J 


5 

'J 
y. 

C 


5 u 

"2 re 

1-1 
* u 

•5 ■■§ 

"^ 2 

ll 

'd "9- 


'^ 
■y 

> 

1> 




oil 


C- 

'-» 

y, 
'U 

'y 
"a 
Z 

u-' 

'-> 

re 
;/: 

c 


z 

■y. 

•0 

y 

u 

"E 


z 

y 


re 

y: 

Z 

'J 

> 

< 

> 

y 

c 



:! 

CIJ 
"y 

re 

-y 

re 
a> 

> 


re 

re 
re 
'J 
y 

re 
"0 
'-» 

5 
^' 

re 
/^ 




'J 

•-» 

■0 
-J 

c 

re 

T3 


T3 

x; 

c 


y 

•y 
-J 

Civ 
y 

y 
'J 

re 

0. 

i 

2 


'-» 
y 

-5 
1 

•-> 


■J 

y 

>. 
y 


CXj 

re 

CIj 

■y 
'J 


'-» 

y 
'J 

'J 


d 

y 
y 

y 

cp 

y 


y 

y 
-3 
*_» 

*_» 

•_» 
y 

> 


y 'X 

"c E 

E '-^ 
c '-* 

— y 
-J — 


'd 

ty 
■0 

re 
c 

'J 



'J 

'J 


^ — 
~ -J 

y , 

^^ 

V-. re 

£ ^ 

■3 5j 

S = 
jj re 


5 

y, 

y 
y 


y 

^, 

'3 

y 

3 

r= 


y 
-y 


1 1 
|l 

— u 

JA 

7. ^ 
.r. ^ 

*" y 

^ z 

— '-/ 

1 J 

< 1 

j^ y 

— ^ 


C/1 

C 
1) 

E 

E 

o 


- ,■? 

i3 "5 
E i^ 
E C 


C/5 

re 

c 
u 

D. 


-i. 
re aj 

g g 


« 
'J 
< 


> 

_re 


5 ^ 

- = 
T= re 

■il 

^ 01 


a. 


c 
r 


" > 

P 


N 

re 



-0 


•J 

'J 


-y 
y. 



> 

re 


2 

"re 


1 

•J 
> 

"re 



re 
C 

'y 


re 
■y: 

Z 

re 

OIj 


■5 

"re 

y 


re 


y 
y. 

•-> 




C 

•J 

•3 


-y 

re 
Z 

>-. 

"3 


y 
'-> 

re 
'J 

-5 


-J 

y 


> 
-y 

'-» 


y. ~ 

> — 

5 - 


c. 

-y 





5 ^ 

"5 "c- 


'_» 


'_> 
y 
■y 


u 


u r 


-J 






'— ' 


S qI 


<r 


!Z 


-J 2: 


^ 


z 


U 


" 


^ 


Z 


2 


" 


^ 


— 


5 


_Z1 


j2_ 


:5 


^ 


2 


'J 






j£i 


r: u 


u 


't; 




u 

Vi 






_1 




Q 


Q 






Q 






a 
















)i^- 


















LU 

d 










3 










ny 






























^ 


























c 




































































T3 *aj 




































































•^ -^ br 




































































2 )3 ° 


— 




D 




> 


> 






> 






; 
















> 


















> 














































































— t. 




































































c F 






































































z 




z 




> 


z 






^ 






^ 
















> 


















> 










3 " 

U 0- 




































































1 u 


















CJ 






^ 


































y. 










18 


IIj 




re 




Zj 


zij 






re 

























:d 


















^ 










o c/1 


















rt 








































5 










Q. 


























































'-» 










>-i 




































































It, 




































































" H 






































































;^ 




D 




> 


> 






3 






^ 
















>■ 


















> 










S i> 




































































.^£ 




































































X 








































































































































||. 


<• 




<; 




a. 


< 













< 
















q 


















< 










H H 










< 






< 






















< 






























i~i 




r^ 




o 


(-1 






r^ 

•* 






oc 
















^ 


















5 














^ 




oc 


<N 






^. 








V, 










oc 


















'^. 










"^ < 










"X 


m' 












3^' 




C^ 










o-' 


















rN 
















— 




re 














re 




'y 
re 












1 






























y: 




s 






n<^ 






C2 







































re 

C- 

u 

H 

U 










Z 

a. 


Z 

c 




Z 

::• 

c 

u 




^ 


z 

1 






Z 

>\ 


CO 










^ 
u Z 




"re 

'c- 
re 

■0 


re 
5> 










re 

s 

E. 


























re 








u 


< 






c 







re 

D. 




"re 

c 


^ 










to 


















c 
















c 




c 














re 
















CO 


















Qi 


































t/: 
















c 


































LJ^ 




re 














U 




re 
Z 














re 
Z 




























u 

^ 


^ 




(~- 




/-s 


C 






f^ 






Q -^ < 














*— J 
















U 










V- 


!5 




1 




2 


2 






S 






SO 




c 








22 
















Q 










Q. 


r-~ 




DC 




_ 








ri 






















^ 


















r- 










CO 


■c 




w 




r^ 


r-- 






r 






f 




r 












r 


















r - 










S 





































































220 



— ^ 

c o 

o — 

E .i2 

2 ;: 



D3 — !M 



CD 



o 01 . — . 

CO n u •^ 



OJ 


-n 


> 


c 




« 








f/i 


CJ 




'/I 




3 

o 


> 


.1. 


1> 




'r\ 




n 


o 


r, 


00 


CC 


c 


a 




n 


<■ 


-a 






u 


<u 






3 


H 




■o 




cz 




r3 






r o 



a> 


lY 


Cl 




r.' 


a 


o 




c 


^ 


OJ 


M 


0/J 


Z 



OH 

o 



r. 


o 




w^ 




u 


w 


rv 












ra 


1) 






■^ 




Z 


a> 


« 






f- 




u 


o 


OJ 


O 



^ m 



1) s 
£ = 



03 2 



W v5 



^i= 



(- 



2i !3 



C ° „ "■ w 



D. -C 






o o- o •= u 



o 

OX) 



.•^ a -a -^ 



o -a 3 5 
I- !^ -' OX} 



-as; 



c 



'U -Q 



8?- 



.— ^ > 



T3 
3 



OXj " 

c a 



U J2 C 
n ". "J 

;i ^ u 



= -a g 

, (u jj o 
t > t/5 



X -s 'J -= -ii 



« 






T3 5> = 



« s -:2 - 

U 3 ?J " 



i S 



^ X) 

■o — 
■^ -a 



1^ o t 



^^ 






5 ^ 

« >. 
-= -a 

CO c 

> <i> 

ij -^ 
w ^-^ 

^' c 

-y o 

"5 "-^ 

n. "^ 



r^-^ — ^oa-^sj 



fy-j CO 

1> P 'J "J 



E 






E 
u 

o 
z 

c 






c/5 

z 



Cl 



C 

O 

■a 

T3 



Oi 



u Z 



2 23 

CO - 

E 3 



O U 



3 § 

C 

" z 

is 



2o 

a 1^ — S r^ 



oj o 
XI o 
■a o 



S *- 



c -S Z c 



.t^ u r; 3 



C/l 



C/1 



— U 

i i 

» CJ 

3 CJ 

XI o 

.-' <= 

5 I 

o 3 



" 2 01) j: 



< 
c 



01) ,^ o 

u .9 3 
V B " 



a> c ^ ■- '-Z 



■5^ 



1- -i! ^ 



5 -e 

E c 

O 

■a 



H 
c 

CO 

Z ^ Ji o -5 -a ^ 



2^ 






> 



co' < 
•5 > 



o 

CJ == 

OB W 



CO ii -S ■ - 



._ '>i ^ CO — 
CJ 3 CO ^^ '.^^ 



o. 
o. 

< 



CO o ■= Ji 



■*.'<•*- 3 



u -o 

Ofl oj 



4J 
-- 3 — -. J= 



•= -^ C 



n. uU 



qj L> 

C O 

■- o 

l> -^ 

— -c 

-3)2 

"E § 

< i 



a "a 
-a 3 



1) x: CO o 



'/■' m r; 



. 2 
■a 3 ^ 



— u. eo — 

x: E T3 XI 

:• e I ^ I 

S i- "U O i- 



o 


CO 


« 


2 


r 


2 


1 




01) 

3 


OJ) 

n 


O 




x: 


o 






CO 


^. 






fH 












•— • 


;« 


01) 

c 


< 


■o 


jj 








OJ) 




c 


CO 


o 



_1J CO 

-3 CO 

> o. 

-= < — 

cri ^ 'yi 

T" -5 .S 

-5 oij 8 

L- ? C^ 

C CO ^ 

CO ^ o 

?^ 3 3 

9- Pi u 



u 



Qi 



O U CO 
>. CO - 



^g 



S2 c 

5 ^ 



^ ~ 



O 01) 

O 3 

ffi :■= 



> x: 

C 3 



■z: -c 






O 

E- 

3 
CL, 



oi 






LL) 

d 



3 •£ O 

o "3 fi 



>- 



3 i- 



> 



.5 u 

Is 

O C/5 



o 
o 
o 



■r- 1) 

1 






< 



— n. 

5^ 



a: 



a. < 



ON 






E 

CO 

Z 



0- 3 
I/-, O 



CD 

3 . 



a 



It 

■^ < 

C- 01) 

Z iS 

D. E 



_<: Z 

ij 



Z 



o 
■a 



o 

CL 



z 



73 
a. 



00 

c 

CO 

f- 



o 
o 
CD 



Q. 
CO 






< Q O 
> S Q 



< 
> 



< 
> 



< 
> 



> 



OS 



> 



221 





-o 






















































•_ 










^^ 


c 


3 


12 

3 








C3 




-3 




Z 






s 








C 


Z . 


■3 






C.' 


re 




> 


y. 

y y. 


- 'A 

y. "3 


iH 






? 

o 


9 

Q. 


c 
o 


O 
'J 






3 

-D 
-J 


'J 


'A, 


-■3 


^ 

X 


s. 


^ 






■^ 




y 




3 






'-i 






^ 






>. 











■/I 

c 
c 




> 




if 


'■J 


'J 


-3 
-3 




x' 


^-* 








\i 






^ '-» 


*-> 






y. 


•t5 


'J 


'J 
y. 


P -y 
7 ^ 










3 
C 
to 


S3 

E 

c 

c 


"a 


E 

> 

-o 


'A, 

c 
■yi 




E. 


re 

y. 


£1. ■ = 

— ■ 9 
o 2 


"J 
9- 


1 

H 


z 


'J 


^ '5 

y ___ 

— -y. 




— 

z 


2 


V 
is 


2 .,; 








'•J 
J-. 

S- 


> 


'J 
o 

;y 

■y. 

■y. 
'_> 

'J 

■y 

-a 


y. 

-a 

re 


2 -I 


"3 "c 


5 

c 


c 
re 

^5 




O 


2 

> 

D 

c/i 

C 

u 


.c 
To 

01) 

1 


> 

c 
'-> 

a. 

Q. 
C3 
<J 


•A, 

u 

03 


i 

Qi 
S3 


> 

.3 


1> 
'-J 

■a 
'J 

'J 
1> 


^ o 


r 

•3 

3 

'J 


•3 

'5 

> 

y, 

"re 


H 
> 
-J 


> 
Z 


"re 

iyi 


o. > 


y. 

N.' 


> 
Z 


'J 
-y. 


■y, 

^» 
y 


£ -i- 

— C3 


'J 

r3 — - 

— 'J 




o 

re 

5^ 


-y 

'-J 

'-» 
<• 


> 


y. 

'd 

-J 

_^ 

'J 

5 
'_» 

"E. 


■3' ~ 
'V' 3 

_3 '-y 


3 '-* 
'J — 

? -£ 


> 
re 

y 
y: 
'_» 

"Ej 


y. 

'J 


ifi 
c 

1 

i 


3 
01) 

,_2 


^ 1 

— « 
'J > 


T3 

_i_ 




c 

U 


>• 


!7 
— 

'J 
y. 


3 


D. 

-yi 

'J 


oil 

3 
O 
>. 


1 ^ 

IB 


"2 
'J 


■y; 


y ^ 
'-» _' 

?. < 

•f s 

•y — 






■y, 

H 


-y 

Z 


^ y 

S^ J. 

Si 

O 

^ > 


2 '-> 

5 re 
a.Z 




3 

o 


y 
-y 


> 


-y- 
'J 




■— re 

3 y-, 

■3 IE 


> 2 


re 

'J ^_^ 
C y 

'-» C 


> 






































-^ 




















u 








Q. 




















o: 






d 




Q 




/-^ 


""; 




,-^ 






_ 










d 






^ 


3 




































^ 










^ 
















Q. 
































































C-- 
































































■O OJ 
































































-= ^ >- 
































































§ s s 




















^ 






>- 




>- 




>~ 


> 




>• 






> 










>- 






> 


































































^t; 
































































c c 
































































5 i> 




















> 






> 




>- 




!>- 


> 




Z 






z 










>■ 






> 


3 >-- 
































































u a. 
































































tential 
cope 




















CO 

S 






- 




« 




^ 


'J 

C3 
O 
O 
O 




eg 






y. 










1 






yt 
*-> 


C (/^ 




































-f 




























c- 




















r^\ 
























'^. 
















tr, 


>x 
































































73 ^ 
































































o c 
































































!0 tl 




















> 






>" 




>- 




^ 


^ 




>- 






>- 










>• 






-^ 


































































































































■^ 
































































i, y 


























^-^ 




o 




■— ' 






;^ 

























u a. 




















< 






_' 




__' 




^' 


< 




^^ 






>— • 










X 

< 






— 


Hp 
























<*' 




< 




-u 
< 




<* 






<;' 














^ 


ii 




















o 






r- 




oc 




— 


OC 




*/■, 




r^, 


__ 










5 






■^ 
























«r, 




— ; 




f"^. 


»A, 




■rj- 




OC 


r^4 










f*i 






>c 


0- < 




















r-f 






z 








-r 


o 




r-". 






rj 










S 






rj 
















































„ 
















^ 






















































































0. 
Z 






< 




Z 

E? 

o 






~ 




"c_ 




Z 

> 
re 


Z 










5^ 






Z 


u 






























Z 


Z 




5 


— 

o 










i 

^ 









z 




















u 






i* 




^^J 






"O 




O - 


- 




C 










U 






S(j 


^ 
1 




















u. 










'J 

:5 






J 




,|z 


t:^ 


■3 
















3 






















E" 






2 








— - 


^ 




3 




^ 










u 






~ 






















c3 






w 




^ 




— . 


^ 




^ 




re 


" 










3 






J^ 






















I 






O 




z 






3 




"« 




2 


" — 










CO 














































£ 

^ 






'5 







































> 






> 




>- 




> 


>■ 




>- 




i^ 












< 






U 


^ 




















^ 






S: 




:S 




:$ 


L^ 




1^ 




u: 


z 










z > 






Z 


% 


















"~~ 












































■■^^ 


Q. 




















•r. 






1 




jC 




^ 


^ 




— 




r J 


•2 










"_ 






r; 


(0 
































































S 




































~ 




~ 




~ 


' 










" 






~ 



-)->-) 



ii u 



C SI 

a — 



3 I- 

" E 

< a 



u c 



:r ^ u 

O C u 



1J o 



u on 

9- c 






— M 

3 O 



2 ?;. C 



3 O 

o -a 

■— « 

o ^ 






c '^ ^ 

"!^ "o t; 

go-" 



■o = 



■= G > .H. 

o ^-H -° 
> " u 3 



"J 

^ I 

3 

ii 15 

t3 C 



=3 'fi 
3 ^- 



Z U H 



H 'f '^ ~ 



■£ Z J: -^ 

^ i» o -u 



t" E S 
t! "5. u 






"Z 



O '-> s 

ir ^ ■ p- 

• T3 '^ 5 

Z c ii ° 

< J3 ^ ^ 



t, 'J CJ 

S 5 w 

u 2 ^ 

CO 3 C 

OJ O" O 

J- ^ ^ 



C/l 



3 ^ - 

= it 



to -a 

o .ii 



i> ^ .^ £^ 
?? ,^ — • 



s 


_o 


c 




ra 


C' 




•-. 




'/■i 






CO 

c 




nn 


a. 






o 


cyo 



on .^ 
S3 "2 






P3 _0 

5) "d 

'J T3 

I § 



3 -^ 

S 2 

" o 

S U 



> 2 

O 3 
— > 



■o y n 



t; u y 



■^ 3 



a -o 

G ^ 

n. N 



u •-- 
C ~ 

o " 

Q. J 

jj :s 



E - E 



o 

D. 



-^ n S 2 



■- ,9 •£ 



■3 o 



L^ 



"c3 Z 



jj cJ 



Z :5 



3 -6 cr 
^ — ? 



u 



£ -5 5 

^ i a' 

5 ^ c 

;'~- i5 '-J 

'" "^ §■ 

OJ c S 

!^ J 

y "" ^ 

f— 3 < 

I ^ « 

^' ■— C/5 

Ca CJ 

i: TT — 

.'J ^ ^ 

., ou h 



^^ ^ 3 

° u t^ 



' c 
o. 

U OJ 

= "5 - 
O .^ = 



CJ "3 

E 



•5 < 



« -c — 



t^ CJ .^ 

-5 -C 
c/1 "O 



1_ </l 
I ° 

■E ^ 
c c 

on 

o 

ij ~ 



c 
Oh a 2 

CJ CJ 

■-01 = 

CJ c ^ 

£ « 2 

O. " 



< i2 2 



CJ 



CJ 



c ■= 5 :-^ 



w 



CJ 



5.2-5 

O VI CJ 
■•7 3 V? 

c 




O 



Z o 



CJ 



jz a ^ 






"5 ^ "2 






olj — 

o ^ 



-= 5 ,0 CJ 3 y 



a. 

CJ ^ 

■' s ■£ 5 

CJ O *^ 3 

1- CJ o CJ 



CJ _ 

CJ E CJ 

~ ,_^ c5 

3 yj 

_CJ j^. -73 

5 H S 

S 3 i= 



5 g - 

'■^ " S 

CJ _c ?^ 



CD 



i: CJ — - 



't- CJ 



Q 



CJ ^ 



c 


m 






U— 


,_, 




OD 


M 


>. 


= 


c 
















> 


2 


^ 


'J 




^ 


, _' 


t/; 


r3 


"J 


P3 








^ 







!« 


.;_; 


r- 




n 






CA 








'H 











^ 


ou 


C 


r; 




1> 






^ ii CJ ^ 3 



•S ^ > CJ C 



= CJ 

C N 

O — 

CJ — 



y 2 



o 
o. . 



T3 r 



T3 

§ 

C 3 ,j 

_^, '-J 

CJ ca 3 

2 E 5i 

ra 'J 

CJ ^ ^ 

XI > •— 

, CJ 
->-. VI -3 
P3 CJ ^ 

E -3 ^ 
CJ -5 ^' 

S o. ^ 
c o;j ^ 



11 



CJ 

E 



O 



•; o X 



— CO 



, — CJ ■_ 



CJ — i 

, CJ 

S 2 

■^ 3 

E o 



CJ Q- 
> £ 



CJ CJ i 
Oij-= 

a ^ ,y, 

■S ra CJ 

« ^ - 

> O O 



o 

e- 



d 



Q 
etc 



Q 



ac 









2 ^ Si 



>- 



>■ 



Si S? 



>■ 



.S CJ 

S 8 

o c^ 



+ 

o 
o 



o 
o 



■r- CJ 

O S 



>- 



3 



o JJ 

CJ O- 
,*- >> 



O 



I 



X 

< 



< 



< 



-g CJ 
TO CJ 

ft. < 



On 
O 



z 



3 
o 



U 



I 
Z 



CJ 

c 

< 



C/5 



a: 
a: 



3 

o 

Ofl 

m 



a: 



Xi 

O 



CQ 

z 

> 



u 
c 
o 



ca 



t: 



< 
Z 



CJ 

o 
o 

-i 
i 

u 



& 
2 

o. 






00 
o 






y' 



z 



223 









^_ 


































*J 












— 












^ 


























J-. 








^ 










: 


■^ -:ii '-» 5 
























r. 










= 


■r. 
'J 




o 


u 

ra 




'.J 




-J 1— 






r 
-» 


'-> -. y — 

^ ^ "= b; 










c 


r<~, ^ 












-= 










^ 


^ 


i ii 


3> 


^ 




■^ 


^ 


u re 






: 


.^ .= ^ J£ o 


O 








■^ 














.2 










— 


— " 


9 '-> 


n 


c 




5* 


^ 


-^ ■^ 






: 


-^ -^ ^ 


= 










3^ 












'J 










^ 


O 


"2 ^ 


■5 






^ 


gj 


G O 






> 


•5 £ = "; £ 










M 


*— ^ 






















-Jj 


^ 


'_> 


u 




~ 


y 


i- ___ 






; 


■ — — - .^ ^ — 


--J 








O 

t: 














o 

5 










■5 


"3 


- J 


< 


Cl 




£ 
O 


'A, 


o "^ 
o -* 








f ^--^ 5 1 


re 
















'J 

J- 
















'J 


|-2 b 

;= s -r, 
— - o 


o 

o 


1> 

> 




y 


•£ 


5 "^ 

-* ■ — 






> 


O ill 

H 5 ^ y. H 
-"= i= S y 


•_» 








a. 

T3 




1J 

S.'' 










i 

u 

V 










'J 


7 
'J 




^ 

•o 


'•^. 

CJ 

o 




■y! 


-a 


■ "5 






i 


s c z 5 'S 


G 






s 


'^ 


— 'J 


3 




^ 

o 

o 






"V 








y. 






5 S 


^ 


re 

c 




o 


y. 


•J ^ 

re y, 
'J — 


y 


■J-- 


> 


Itpi 








-3 




c ^ 

'-> 'Si 

'^ z' 

w r3 






'J 






C 












o 
•-> 


-_) y; 


IE 

re 


'J 


■y^ 




'J 


1 i 


> 


z 


; 


o -3 ~ 5 - 

c 3 a: = y 

1 N = | 

' z. -^ — 

=P^ .■= = y 


■-* 

y 

*-* 






o 




C > 


U 








= 








"2 


:^ 


S-H 


c. 


c. 


-J 
■J-. 


o 


•-> 


"5 — 


o 


re 


> 


= O .^ _;; ji 




C 
O 

u 


^ 

o 


O. 

z 


o 2 

•j — 

. — o 

if -o 


i2 




2 

'J 
















'J 
<- 


■j: 

'J 


^ 'J 

i 5- 






> 
•J 

'J 




'J 




•| 


"5 


> 

i 


o -3 n t "^ 

y y ?^ ,0 r- 
i V=3< 


•-> 
< 




«J 
























































t/1 


























































o 












/-> 






™ 






















: • ; 




















—- 




_1 






^ 






^ 






















d 








— 




































































^ 

























































-a u 


























































•= ^ t 


























































§ ^ 2 


>- 




D 






>- 






>- 








yf 






z 








~. 




z 




> 




z 






1 "^ S 


























































>% ^. . 
























































c r 




























































z 




Z 






>■ 






>- 








y^ 






z 








= 




>- 




Z 




z 






D 1- 


























































u «^ 


























































^ I. 












•J 


































y. 










?: ci. 












:3 




































U 












— 




ID 






O 






^ 






















^ 








— 




































































s 














































-r 










>* 


















































































































■« ■— 


























































'J F 




























































>- 




D 






>- 






>■ 














z 












^ 




^ 




z 






2 u 


























































?£ 


























































"T 
















































































■^ 




























































































!> aj 


^ 




















^^ 


















— 








^ 










^H* 


^ 




< 






< 






< 






















<■ 








^ 










Si 

- < 


-t 

■X, 




C 




o 




- 


- 


oc 






-r 


















o- 




^ 




tT 




oc 
























y: 










































« 






























~~' 


























01) 


















































*^ 


u 
Z 


s 

z 

c 




g 

c 
« 

^ a. 




z 

> 


z 

c' 


y 
Z 

y. 


z 


'J 








y 






z 

CO 

< 












J' 

z 




J' 

y 

y 
-i 

:5 




z 




z 

ly, 

re 




2 

y. 

E 




5,2 

E 




c/1 
3 


> 


y 

y. 




'J 




o 




O 
















S 

D 




y 




•y 




1 

c 
o 




•-> 










•^ 


cc 




V 




7 




>' 














l> 




J^. 




y^ 








"^ 


"J 
























































Jo 


<■ 




< 




_J 


__ 


y 


y 


y _ ^ 




a£ 




J„ 






lU 








. — 




< 




<" 




<" 




< 


^> 




o 




< 


< 


^ 


^ 


^ < ^ 








2 






s 








> 




2 




s 




^ 




I> 


«: 


































~"~" 


■""" 


"""■ 




















§■ 


r) 












(~1 


r J 


OC 








_ 






^ 1 








-T 




r- 




3C 




— 




r*^, 


s 










~" 


~ 


~ 


"" 


" 









































224 



c 
u 

E 
B 
o 

_y_ 



■a 

c 
u 

c 



c 

o 

U 



f 


>> 




h- 


1J 






^ 










■a 


^ 


oil 




fl> 




C 


^ 


— 


ij 


a^ 


L> 


J2 


5 


ti; 


O 


c 




-. 


« 




S 




M 


0. 


O 


CS 


J= 


CO 


^^ 








q> 


■o 


c 


c 




v^ 


> 


o 
c 


c 



«_ (Q u 



«/~i 


X* 


C3 


c 




C 










o 


u 




un 




CA 




o 




^ 


OiJ 


1J 


U 


OIJ 










3 


^ 


:S 


O 


> 


X 
























O 


"TJ 












c 


•^ 




o 


-n 


•^ 


c 


OJ 


\i 






a: 



o 



2 i 



c 


? 


m 


'ri 


c 


'J 


on 


D 


c 


Vj 






r^ 


fY 


rt 




^ 


a 






^r 


« 






'd 


U 


O 


■o 








rt 












tn 






-o 


3 


u 


« 



u 



o 



S ^ - 

" I ^ 



u 


p 


c 








o 




Xi 




U 



















o 
c 


nj 


C/1 
1> 


11 








o 


c 
<• 


O 


i> 


■o 




on 












a 


n. 








« 


h- 




i> 


7 


1 








f— 






TT 




o 


=; 










« 


o 


o 


1) 


<r 


u. 


ai 


^ 


1> 


o 


75 




a. 


c 

o 


-5 










1> 


<u 


•n 


C; 


-a 


.£ 


T3 


















'/^ 


v~ 




















'C- 


■^ 


— 




+ 

!/0 




-a 
c 


rt 


li- 


-a 


, , 


/, 


ra 




c 




n 


•u 


'^^ 


o 








o 


— 


s, 


O 


o 


^> 


fn 




■a 






£. 




a 


ra 


^ 


o 


r/^ 


o 


X 


■n 


a. 


^ 



cC :: c-a: 



m 





c 






rt 




rt 




*^ 


G 


c 


i 




cd 


r, 




> 


i-* 


o 
c 


o 


.!<: 


VI 


o 


f?l 






D. 


^, 






« 


(>-l 


"t^ 






c 

o 



w "O 



C 3 

Oil ti 

i2 S 



'J _ 



^ a 2 c 
S •= *= ^ 

= " n. on 






cr: 



u 



o 



a. •= 



- D 

-o — ■ 

^ — ' 

c ^ 

n. a. 



5b °' ^ 



J> = s 

O C; 'yi 

3 ^ OJ 

U — (V 
"'■ o 



on 






o 


^ 


■« 


o 


< 




-^ 


c 


^ 


t: 




= 


« 


c 





5 1. t; 7^ - 






o 



> ii 2 - 

■^ ^ ^j ^ 

cj £ a-> ?r 

^ ^ c/i 5 



3 "2 

la 



u 



o 



o 



c 
a 


o 
r 


^ 


c 
E 


'•J 


1J 


V) 








>-i 


s: 










vj 


^, 


cz 


on 


'■J 


r" 






^ 


•^ 


c 


3 

O 






'J 




f3 


^ 


^ 


Q 


'-) 




v 




a. 


•n 




^ 










3 


ra 


^ 


c 




O 


'r 


. . 


O 


oj 


Q. 




<r 


fyi 


n> 




13 


Cu 


on 
S. 


2 


u 




o 


01 


















5 
"J 


o 




ra 


o 

o 

-J 


> 




ra 

CL. 


xi 




It 




■o 


X 


?■ 




- 




',^ 


y^ 


£ 




'•; 


O 



^ < 



0.-0 

on « 



,3J 



y « - 



^ i; c '-> 

^•:! 3 

= c Si „ -^ 

o E 73 y =y 

^ i^ i -S •" 

,1 o 3 ^ J 

— £ c5 0.0 



5 rt c -t: 

U ^ S 3 

_ Cl r- « 

.£ > ^ r 

"^ r"^ Cl. w 



c r3 ii '^ 



•a 


-rj 


'jn 


P3 














O. 




>' 




OJ 


o 


-a 


■^ 



■ CQ 



D. 

c o 



■^ ■y n 



O 



d 



Di 



LU 



Q 
ai 



5 !a ^ 



>■ 



> 



> 



n '6 
u u 

fc '- 



<L> 



o. 
o 



o 



O 

o 



3 



S " 



> 



>• 



3 



>- 



•1> 


u 


s> 


a. 




>, 


H 


H 


^ 


«n 


i3 




G- 


< 



< 



c 

a 



a 
< 



< 



m 
^ 



z 



2 

O. 



Z 

■a 
< 



-o 
o 
U 



U 



c 

o 



£ 

o 

h- 



on ^ 
= Z 



•J x: 

o 



■J o- 



<U o 



a. 
m 
z 



Qi 









60 
00 



z 

00 

3 



u 



oo 



o 



cc 



a. 

o 
U 



< 
> 



< 
> 



o 






225 









y. 




op 










































<r o 2i 






r. 




b 




> 


^ 


-3 






































± S Z re 

z ^ ^- 1 






o 
<■ 




"J 


^ 


■y. 




o 
o 






































:::■- b 1 










j^ 


























































rs 


'-> 


'~* 


r 






































^z 1=^ 












22 


,1^ 


C- 


■^ 






































1 = z 1 






'-J 




T3 


y! 


V. 










































II -1 ^ ^^> 


ti 


i ! ^. ^ 

^ ^ -^ ■ 


■y 








5 


S 




^r* 








































y. 




,-- 01) 




'-> 


O 


y 

o 


_<: 


^;> 






































Z c « = 


C 
CI. 




























































O TZ 

'i — 




-3 

'J 


X 


O 

g 


M 


^ 






































IH ^ — ^ 

-_( ~ cs u 

^ ^ z ^ 






'— E 




•a 


O 

o 


'J 


•a 


^3 








































S 






E 


'J 

5 


■5 

a 
■a 


•J 

c 

> 








































2 - -if Z 

5 Z « II 


1 
ii_ 

■^6 




c ""* 




^ 

v- 


OIj 


CTj 








































z 5 «5 




E 
o 
O 






CJj 


> 


•y. 


"t3 






































-i _^ i :!. 


T 


^5 






■y 


■y. 


■u 








































ilz ^ 


•~ 






^ 


r: 


■o 


u 


CJ^ 








































t: 












































































































u 
























































O 




















































1 z i = 


y 


3 


— 








^ 










































1 i' ^. S 


"u 


ft. 




















































— -* 












































>- 










— — C II 


< 


ould 

ark 

;tore'' 


>■ 








>- 




































< 

> 




z 


5 Z fc 2i 

lis- 




























_ 


< 












3 


> 




< 


^ 


z 


li ■- - c 


II 




















< 


< 


<; 


<^ 


2 

H 

U 


> 
Z 


—1 


z 


z 


< 

0. 


< 


< 


< 

> 




Z 

z 


>- 


= 1^ = 1 


< 


















.i>c:; 


















J5 


s 


s 


s 


i 












> 


> 




2 




< 


z 1^^ 


c- 




z 








Z 






























z 


O 






> 


^-^ 




y 


o ^ 










































~ 










iiJI 


_^ 












































^ 










II. 


"H 










o 






— u 


'J 








o 












































C 

2 
re 

3 

-o 

II 


Potent 
Scop 


r 1 








o 










































IE p. 

~ c re > 
Z 2 z 5 

II re II -3 


>i 


















































1 ^ 




















































a z ^r csc 


LU 




>- 








> 










































?^ii 


=' 


.^£ 


















E 
m 
Z 






J-. 


> 


7 
> 

y. 

'_» 


z 

9 


— 


y 

z 

yj 


-• 

Z 


Z 


Z 




z 

■y. 


Z 


z 

> 


z 


- _) Z -=: 
^ = .. re 
2 Z n Z 

Ills 

:i oj E ^ 

1 5 ^ 1 


C 

o 
T 


Tree 
Type 






r) 


to o 

1^ < 


r 1 
rl 




o 


















u 


X 


y 
Z 

c 






z 

■o 
c 






2£ 


^ 
£ 


r. 


'J 

:y: 

< 




y 






'S 

y 

*_> 

'J 


y 
'J 

'J 


CI) 


--> 


O 
Z 






•-> 


u 

> 


•^ 


c^ i J i 

I- z _• 1 


i 


c3 


X 








u 






^ 






^ 




_u: 


3 






Z 


^^ 


' 




r3 










= 11 5 5 

^ ^ = re 
re Z o II 


r*^ ^ 


Z 


c^ 


Iji 






cc 












^ 




rc 


_g 














'-> 






^ 




^ c 


•i 


5 














■tf 






_? 






U 














- 










s = 




z 


""■ 






^ 






I/: 
•< 

a: 












^ 
























ill! 


i^ II 
II z 






^ 












■X 




































o re « J 






















































X - Z - 


u 
















>^ 




































k— r^ 




w 


o 


< 






^ 










































!i ii n - 


> II 

1> ,-) 


(Jo 


;/; 


^-> 






< 






■^ 


« 


































:!i - 2 5 


















= 




'/•, 


r 


>: 


r") 


-r 


2, 


'I- 


5 




yr , 


rj 


i 


s 


# 


- 


— 


= lz 5 
- .2 — ' re 


o. 


o 


o 






«/-, 






« 
S 


— 


oj 






r-j 






<* 




































Iff 


> 



226 



Steiner, K. C. and Carlson, J. E, eds. 2006. Restoration of American Chestnut To Forest Lands - Proceedings of a 
Conference and Workshop. May 4-6, 2004, The North Carolina Arboretum. Natural Resources Report 
NPS/NCR/CUE/NRR - 2006/001, National Park Service. Washington, DC. 

^ ^ SUMMARY OF FACILITATED WORKSHOP ON 

RESTORATION OF CHESTNUT TO NATIONAL PARK SYSTEM LANDS 

:^ i; James C. Finley and Kim C. Steiner 

School of Forest Resources, The Pennsylvania State University, 
University Park, PA 16802 USA (fj4@psu.edu) 



Part of the goal of this conference was to develop a common understanding among participants of 
directions for American chestnut restoration programs on lands managed by the National Park Service 
(NPS), the principal sponsor of the conference. To achieve that goal, the third day was devoted to a 
facilitated workshop on this topic. Specifically, participants performed a guided SWOT (strengths, 
weaknesses, opportunities, and threats) analysis of the merits of NPS participation in the restoration of 
American chestnut given available technologies (including genetic technologies), existing knowledge of 
the ecological context of chestnut restoration, and likely regulatory and social dimensions of pursuing 
restoration using technologies available now or in the foreseeable future. 

Strengths and weaknesses are internal issues, in this case the advantages or disadvantages that NPS would 
have as a participant or player in chestnut restoration. For these two discussion points, the focus was the 
organization itself relative to its role in reintroducing the chestnut. Opportunities and threats were defined 
as existing or potential conditions, external to the NPS. that might influence decisions to participate in 
restoration of chestnut. Opportunities and threats could include economic, ecological, and social 
influences, including the current state of knowledge and technology regarding chestnut restoration. 
Depending upon context, the same item can be regarded as either an opportunity or a threat. 

Each participant was assigned to one of four groups for preliminary discussion of issues. Forty-two 
persons participated in this e.xercise (Table 1 ). Each group met in turn to consider the strengths, 
weaknesses, opportunities, and threats of NPS participation in the restoration of the American chestnut, 
and all participants reassembled after each session to report group lists. Following this, the lists were 
simplified by combining similar items, and each participant was asked to cast three votes for the most 
important item(s) in each of the four SWOT categories (a participant could place one vote on three items 
or use two or all three votes for a single item). The results of the voting are shown in Tables 2 through 5 
for items that received two or more votes. 

NPS strengths related to the restoration of American chestnut (Table 2) are predominantly associated with 
the size and history of the agency and its archival resources, reputation and appeal to the general public, 
land ownership and tenure, and effectiveness at educating the public. The preservationist mission of the 
NPS is also seen as an advantage in this context, and the federal "red tape" surrounding the 
implementation of NEPA (National Environmental Policy Act) regulations is seen as a safeguard against 
mistakes. The diversity of parks and their missions, and the relative autonomy of individual parks, were 
highlighted repeatedly as unusual for a large federal agency and regarded as potential strengths. 

But many of the strengths of NPS were also cited as weaknesses when viewed in another light or by 
different participants (Table 3). in fact, the last-cited strength - park autonomy and diversity of missions 
- received the most votes as a weakness because autonomy can lead to poor coordination and 
inconsistency in policy, practice, and priority. Not surprisingly, insufficient or inconsistent funding (both 
intramural and extramural) received a large number of votes as a weakness (when is this not true?). Most 
of the remaining votes went to the existence of policies and cultural attitudes that stand in the way of 
applying new (or old) technologies to chestnut restoration. There was a lot of support for the idea that the 



227 



lack of a long-term NPS plan or policy is an impediment to pursuing initiatives like chestnut restoration. 
Of course, the principal reason for the workshop was to help formulate plans and policy. ._..->, 

Chestnut restoration seemed to be viewed as a good thing by most participants (Table 4). The potential : - 
emergence of new technologies and blight-resistant varieties was seen as an opportunity thai the NPS, 
given its unique and significant role as a major land steward within the American chestnut native range, 
should pursue. Chestnut restoration was seen as probably ecologically beneficial to the forests managed 
by NPS. but participation in restoration was also seen as beneficial to the agency itself through the 
engendcrment of public support, the furtherance of useful partnerships outside the agency, the 
enhancement of educational programs, and (more obliquely) the opportunity to sharpen NPS policy by 
focusing on a model species. 

Most threats emphasized by participants (Table 5) seemed to arise from what could be called "the prudent 
exercise of extreme caution" rather than from actual knowledge of risks. Participants focused on 
unknown ecosystem effects arising directly from chestnut restoration (whether using transformed or 
hybrid material), the possibility of failure due to breakdown of resistance or the emergence of new 
diseases, and the unknown consequences of the changes that have occurred in the forest over the past 70 
years and may occur over the next centur> . Also cited was the possibilit\ that NEPA and other legal and 
regulatory issues could lead to a quagmire before restoration could even begin. 

Workshop participants appeared to believe that NPS should at least articulate policy and at best actively 
participate in the restoration of chestnut in view of the NPS mission and its ownership of significant tracts 
within the original chestnut habitat, and also the fact that technologies may soon be available to actually 
achieve this long-sought goal. Clear policy or at least policy guidelines would alleviate the principal NPS 
weakness identified by participants - an inconsistency in policies and priorities among parks. 
SurprisingK, a significant number of participants (influential in the vote tally) seemed less interested in 
the potential ecological benefits of restoring this ke\ stone species than in the unknown risks associated 
with non-native genetic material and possible perturbation to forest ecosystems that are, if not stable, at 
least reasonably robust after recovering from the loss of chestnut. This outcome seems to reflect a "desire 
to understand everything before doing anything" as identified in Table 3. The consensus appeared to be 
that chestnut restoration is a "high-gain" pursuit. Whether it is a "low-risk, high-gain" or a "high-risk, 
high-gain" pursuit remains unresolved by this workshop. The other papers in these proceedings ma\ help 
answer that question. 

Table 1 . Participants in the facilitated workshop (M indicates the designated moderator). 



Group 1 


Group 2 


Group 3 


Group 4 


Kim Steiner, M 


John Carlson, IVI 


Paul Sisco. M 


Tim Phelps. M 


Bill Lellis 


Jim Sherald 


Ray Albright 


John Karish 


Jennv Beeler 


Brian Carlstrom 


Jennifer Lee 


Kristen Allen 


Larr>' Hi la ire 


John Perez 


Jennifer Hewitt 


Michele Webber 


Becky Loncosky 


James Voigt 


Chris McNeilly 


Tom Blount 


Mark DePoy 


Greg Eckert 


Mar\ Willeford Bair 


Kent Schwarzkopf 


Paul ficrrang 


Joe James 


Ries Collier 


Matt Diskin 


Tom Kubisiak 


Fred Hebard 


Scott Schlarbaum 


Sharon Friedman 


Phil Pritchard 


Albert Meier 


Songlin Fei 


Benji Comett 


Bill Powell 


Peter Gould 


John Bellemore 


Sara Fitzsimmons 


Dave Loftis 


Will McWilliams 







228 



Table 2. "Strength" items relating to NPS involvement in chestnut restoration that received multiple 
votes from participants. 



Votes 


Description 


14 


The NPS has a great deal of ecological data and information regarding the natural resources of 
national parks. 


13 


The NPS has a great deal of appeal and goodwill within the American public. 


9 


The NPS manages a large number of parks throughout the natural range of American chestnut. 


9 


NPS land ownership is long-term. 


7 


The NPS does well at outreach and public education. 


7 


Much of what remains of the natural genetic diversity of American chestnut is represented 
within national park lands. 


7 


NEPA implementation in the NPS provides a deliberative mechanism for environmental 
decision-making. 


5 


The diversity of parks and their missions within the NPS offers a variety of venues by which 
restoration can be approached. 


5 


Restoration and control of exotic species (like chestnut blight) have broad public appeal. 


4 


NPS lands are refugia against commercial exploitation. 


4 


The relative autonomy of individual units creates opportunities for many approaches and 
experimentation. 



Table 3. "Weakness"' items relating to NPS involvement in chestnut restoration that received multiple 
votes from participants. 



Votes 


Description 


20 


Individual units tend to operate rather independently, with resulting inconsistency, poor 
coordination, and variation in policy and priorities. 


13 


Funding is inconsistent and budgets and staffs tend to be small. 


8 


Competitive funding is uncertain and tends to support only projects in the 1-3 year range. 


8 


The NPS has no clear long-term plan that addresses issues like chestnut restoration. 


7 


Existing policies may stand in the way. 


6 


Private, proprietary rights to blight-resistant chestnut material could be an issue with the NPS. 


5 


The NPS can have a preservationist mindset that slows action, especially the desire to 
understand everything before doing anything. 


4 


Deer control is usually inadequate within national parks, and this could be an impediment to 
planting trees. 


4 


There is a perception within the NPS that manipulation is against regulations or at least 
discouraged. 


3 


Our inability to define "natural" stands in the way of defining management goals. 


3 


The NPS has had a history of poor cooperation with the USPS. 


2 


Staff and expertise are lacking for large-scale, landscape restoration. 


2 


NPS planning tends to be lengthy and cumbersome. 



229 



Table 4. "Opportunity" items relating to NPS involvement in chestnut restoration that received multiple 
votes from participants. _ ^-^ 



Votes 


Description 


18 


New technologies and blight-resistant chestnut seed to enable restoration may be available 
soon. 


16 


As a large landowner, NPS can be a major participant in large-scale restoration. 


8 


Public (e.g., USPS) and private (e.g., landowners) cooperators are available for this work and 
offer opportunities not only for leveraging resources but also extending partnerships. 


7 


Chestnut restoration can enhance ecological integrity and stabilit> of parks. 


7 


Now is a good time and chestnut is a good species for developing a policy model for species 
restoration. 


7 


Participation in chestnut restoration would play on public enthusiasm and engender support 
for the NPS. 


6 


Chestnut was a keystone species, and its restoration would have secondarv benefits. 


4 


The chestnut stor> is a good vehicle for education about natural resources, exotics, and many 
other issues. 


3 


There is plenty of time to act on this issue so we can do so with all due deliberation. 


3 


Chestnut could be a replacement for other declining species. 


2 


Chestnut has characteristics duplicated by no other tree species, and it may be more adaptive 
to possible future changes in the environment such as climate change and air pollution. 


T 


Chestnut restoration has a social as well as a biological dimension. 


2 


Chestnut restoration could be the environmental success stor> of the 21st centurv. 



Table 5. "1 hreaf" items relating to NPS involvement in chestnut restoration that received multiple votes 
from participants. 



Votes 


Description 


21 


Restoration could have unknown negative ecosystem effects such as the inadvertent spread of 
pests, displacement of species, and even displacement of native chestnut. 


14 


Some may object to the use of GMOs or backcross hybrid trees to restore a native species. 


11 


Resistance could break down in time as the fungal pathogen evolves. 


10 


NEPA and other legal and regulatory issues could lead to a quagmire. 


7 


Other diseases and pests could ruin our efforts to restore chestnut, and other exotic pests could 
prove to be as ecologically devastating as chestnut blight. 


6 


There is a possibility of failure following expensive and widely publicized efforts. 


5 


Environmental changes since the 1930's (sudden oak death, air pollution, climate change, etc.) 
ma> mean that the forests are not the same as what chestnut disappeared from. 


4 


NPS has no control over adjacent lands. 


3 


It is difficult to plan for a moving target (state of the forest ecosystem) 100 years from now. 


3 


Institutional fatigue can be disastrous to long-term projects. 


2 


Our scientific knowledge of this subject is incomplete. 



230 



.,^. Service 

ment of the Interior 

National Capital Region 
Center for Urban Ecology 



National Capital Region 
Center for Urban Ecology 
4598 MacArthur Blvd., N.W. 
Washington, DC 20007-4227 

www.nps.gov/cue 



The Pennsylvania State University 

School of Forest Resources 
The Pennsylvania State University 
1 17 Forest Resources Building 
University Park, PA 16802 



PENN State 



pa^ 



Southern Appalachian Mountains 
Cooperative Ecosystem Studies Unit 

Department of Forestry, Wildlife and Fisheries 
The University of Tennessee 
274 Ellington Hall PS Bidg. 
Knoxville, TN 37996 



'""> "; 




SA-CESU 

Southern Appalachian 
Cooperative Ecosystem 

Studies Unit 



Chesapeake Watershed 
Cooperative Ecosystem Studies Unit 

Center for Environnnental Studies 
Appalachian Laboratory 
University of Maryland 
301 Braddock Road 
Frostburg, MD 21532 



^^^^^ptAKEW^f^^^^ 




ICOSVSMM SIUDII\ UNII 



EXPERIENCE YOUR AMERICA 



TM